The Kainoki Treebank front page
This manual details an annotation scheme for parsing contemporary Japanese. Syntactic structure is represented with labelled parentheses adapting the format of the Penn Treebank (Bies et al. 1995). More particularly, the Penn Historical Corpora scheme (Santorini 2010) has informed the ‘look’ of the annotation. This includes:
Annotation practice is also heavily inspired by the SUSANNE scheme (Sampson 1995) regarding adoption of construction analysis, and most in particular for the methods of representing complex expressions with subordination and/or coordination. The annotation also contains plenty that is innovative, including information to resolve both inter and intra sentential anaphoric dependencies from a discourse perspective.
Annotation is carried out with a general text editor on text files.
Annotation practice strives first for observational adequacy. The aim is to present a consistent linguistic analysis for each attestation of an identifiable linguistic relation or process. Relations and processes are treated uniformly as much as possible, and their treatments are detailed in this documentation. Setting aside the issue of whether the system of description is theoretically correct, the practice is to render lexical and functional items, parts of speech, constituents of various categories and functions, and constructions defined by combinations of properties, in such ways as to be unambiguously identifiable. This documentation sets out basic principles both for the annotator (assigning segmentations, tags, and structural positions) and for the user (searching for classes of items, categories, and relationships between these). Examples of suitable tools for searching the parsed data without requiring any modification to the data include: CorpusSearch (Randall 2009) and Tregex (Levy, and Andrew 2006).
The current annotation also aims to offer syntactic analysis that can serve as a base for the subsequent generation of meaning representations using the methods of Treebank Semantics (Butler 2015). To this end, extra disambiguation information is added to feed the calculation of semantic analyses from the syntactic annotation. One example of this is seen in the different specifications of clause linkage (i.e., different types of non-final clauses). The annotation has the -SCON tag extension to identify clauses integrated with subordinate conjunction. Subordinate clause status influences the distribution of empty subject positions within such clauses and the antecedence relationships these positions have with respect to upstairs arguments (according to an antecedent calculation called ‘control’). These cases are contrasted with coordinate clause linkage, also identified with a tag extension: -CONJ (coordinating conjunction). Status as a coordinate clause influences the distribution of arguments shared between that clause and one or more other clauses (according to an antecedent calculation called ‘Across the Board extraction’ (ATB)).
With these calculations in place, most antecedent relations in Japanese can be accurately determined without resorting to annotation with overt indexing, provided that the distinction between subordination and coordination is properly annotated. The practice provides a robust basis for calculating semantic dependencies, a simplified annotation scheme that is descriptively adequate, and a set of constraints on the distribution of null positions which have interesting consequences with respect to, for example, the placement of null elements.
The purpose of the rest of this chapter is to introduce broad aspects of the annotation scheme, while subsequent chapters will focus on particular topics.
Tag labels are either word class tags (e.g., N=noun, P=particle, ADV=adverb), phrase level categories with minimally a basic label to indicate the form of the constituent (e.g., NP=noun phrase, PP=particle phrase, ADVP=adverb phrase), or clause level categories (e.g., IP-MAT=matrix clause, CP-FINAL=matrix clause with sentence final particle(s)). Frequently, label extensions (separated by a hyphen) are added to labels to indicate function (e.g., P-ROLE=role particle, NP-SBJ=subject noun phrase). In most cases there is one label extension, but more is possible (e.g., CP-QUE-OB1= question used as object, IP-ADV-SCON-CND=conditional clause).
All tag labels used in this schema are listed below in Tables 1.1–1.5.
| ADJI-MD | modal イ-adjective |
| ADJI | イ-adjective |
| ADJN-MD | modal ナ-adjective |
| ADJN | ナ-adjective |
| ADV | adverb |
| AXD | auxiliary verb, past tense |
| AX | auxiliary verb (including copula) |
| CL | classifier |
| CONJ | coordinating conjunction |
| D | determiner |
| FN | formal noun |
| FW | foreign word |
| INTJ | interjection |
| MD | modal element |
| NEG | negation |
| N-MENTION | mentioned expression |
| NPR | proper noun |
| N | noun |
| NUM | numeral |
| PASS2 | indirect passive |
| PASS | direct passive |
| P-COMP | complementizer particle |
| P-CONN | conjunctional particle |
| P-FINAL | final particle |
| P-INTJ | interjectional particle |
| PNL | prenominal |
| P-OPTR | toritate particle |
| P-ROLE | role particle |
| PRO | pronoun |
| PUL | left bracket |
| PUQ | quote |
| PUR | right bracket |
| PU | punctuation |
| Q | quantifier |
| SYM | symbol |
| VB0 | light verb |
| VB2 | secondary verb |
| VB | verb (or verb stem) |
| WADV | indeterminate adverb |
| WD | indeterminate determiner |
| WNUM | indeterminate numeral |
| WPRO | indeterminate pronoun |
| ADVP | adverb phrase |
| CONJP | conjunction phrase |
| INTJP | interjection phrase |
| NML | intermediate nominal layer |
| NP | noun phrase |
| NUMCLP | numeral-classifier phrase |
| PNLP | prenominal phrase |
| PP | particle phrase |
| PRN | parenthetical |
| CP-EXL | exclamative |
| CP-FINAL | projection for sentence final particle)projection for sentence final particle |
| CP-IMP | imperative |
| CP-QUE | question (direct or indirect) |
| CP-THT | complementizer clause |
| FRAG | fragment |
| IP-ADV | adverbial clause |
| IP-EMB | gapless noun-modifying clause |
| IP-MAT | matrix clause |
| IP-NMZ | nominalized clause |
| IP-REL | relative clause |
| IP-SMC | small clause |
| IP-SUB | clause under CP layer |
| multi-sentence | multiple sentence |
| -ADV | unspecified adverbial function |
| -CMPL | used with complement function |
| -CND | used as conditional |
| -CNT | used with continuative function |
| -CONJ | conjunct of coordination |
| -CZZ | used with causee function |
| -DOB1 | used with derived primary object function |
| -DSBJ | used with derived subject function |
| -LGS | logical subject |
| -LOC | used with locational function |
| -MSR | used with measurement function |
| -OB1 | used as primary object |
| -OB2 | used as secondary object |
| -POS | used with possessive function |
| -PRD | used as predicate |
| -PRP | used with purposive function |
| -SBJ | used as subject |
| -SBJ2 | used as secondary subject |
| -SCON | subordinate element of subordinate conjunction |
| -TMP | used with temporal function |
| -TPC | topic |
| -VOC | vocative |
| FS | false start |
| LS | list item |
| LST | list |
| META | metadata |
The annotation scheme represents syntactic structure with labelled parentheses. All open parentheses have an associated label, representing non-terminal nodes in a tree.
All corpus files are formatted to be compatible with the CorpusSearch program (Randall 2009; http://corpussearch.sourceforge.net/CS-manual/YourCorpus.html). Under CorpusSearch format, every tree has a ‘wrapper’. A wrapper is a pair of unlabelled parentheses surrounding the tree content together with an ID node. Furthermore, to support the on-line corpus interface, the ID node must contain a string that begins with a number corresponding to the number of the tree in the file which is followed by an underscore (‘_’) and in turn followed by the name of the file (minus the ‘.psd’ extension). Any other string material may then follow the filename, provided there is an intervening semicolon (‘;’).
Words are terminal nodes that project a node with a word class label. Null elements (zero pronouns, traces, etc.) that aren't in the text itself but are introduced to describe structure are placed directly under phrase-level nodes, and so have no word class node projection. Null elements have terminal nodes that begin and end with ‘*’. See sections 1.5 and 1.4 for details about null elements. Terminal elements that are words and so under word class nodes are surrounded by ‘*’ when they are insertions made to the data from the stages of annotation, as in the recovery of content for instances of particle omission (see section 2.14).
The policy for segmentation and part-of-speech labelling follows the principle of using terminal nodes that are as large as possible, but not so large as to incorporate into purely lexical elements other elements with functional roles. Such a policy corresponds closely with the LUW (Long Unit Word) standard of the Corpus of Spontaneous Japanese (CSJ; Maekawa 2003) and the Balanced Corpus of Contemporary Written Japanese (BCCWJ; Maekawa et al. 2014). A LUW (Long Unit Word) is composed of at least one SUW (single-morpheme Short Unit Word), which corresponds to an entry in the UniDic dictionary (Den et al. 2008), but complex LUWs containing more than one SUW are common.
The chunking obtained with LUW analysis is not limited to complex nominal expressions or complex predicates: Heterogeneous strings that appear to have undergone grammaticalisation (e.g., some formal noun/particle pairs, some complex modal expressions, etc.) are chunked as well. Complex LUWs are usually incorporated just as single segments. For example, numerals are analysed digit-by-digit into component SUWs, but the parsed annotation chunks these into a single segment according to the LUW containing them.
While the LUW chunking of the BCCWJ and CSJ is intended to identify units with significance in syntax, the information is not always rich enough to generate immediate constituency trees approaching descriptive adequacy for syntax. Depending on the circumstances, SUW analysis may not be sufficiently fine grained, making it necessary to split the SUW into more than one segment. For example, the volitional form of a verb may be analysed into a combining stem and a volitional morpheme: (VB 結ぼ) (AX う). Conversely, a series of LUWs may need to be concatenated under one terminal node label (an instance of further chunking). For example, a series of LUWs that together form a complex proper noun may be chunked into one segment. Furthermore, some finer distinctions in morphological analysis that have no consequence for syntax (e.g., the distinction between personal names and place names) are sometimes ignored, while other distinctions deemed important (e.g., sorting instances of items that share the same phonological form into more than one part of speech according to their grammatical function) are introduced. This is a consequence of aiming to expose the basic functional structure of the language, while keeping the structure fairly flat and easily searchable.
The policy is to chunk as large as possible in the automatic parse, and this is the form of segments that annotators initially see. But when there is clearly some constituency in a string that must be expressed by structure, or when there is a need to indicate the semantic effects of structure, chunking may have to be undone. To exemplify the former situation, consider the morpheme 中 (chuu). Following a ‘verbal noun’ like 旅行, 中 is analysed as a nominalising suffix by UniDic and grouped together with a preceding string to form a LUW. This analysis is appropriate for a situation in which 旅行 takes a noun modifier, as in (1.1).
彼 は 旅行中 にして 、 病 に 倒れ た 。
But in a different context 旅行 may appear with arguments or adverbial elements, and 中 is better analysed as a noun, as in (1.2).
佐藤さん は 海外 を 旅行 中 だ 。
Another scenario where chunking may have to be undone involves complex particles. UniDic chunks many verb-particle combinations into complex particles. Note how in the string にしたがって in (1.3) the part corresponding to the verb 従う has been bleached of its semantics.
「 地 は 生き物 を 種類 にしたがって いだせ 。
While the UniDic analysis is frequently correct, there still appear instances where the part corresponding to the verb 従う should be treated as a full-fledged verb, as in (1.4).
モーセ は 主 の 命 に したがっ て 、 パラン の 荒野 から 彼ら を つかわし た 。
In a scenario such as (1.4), the annotator must split the segment and relabel its parts as necessary.
Phrase structure is roughly describable as in (1.5):
Words that project phrases are phrase heads (e.g., N, P, ADV, etc.). Following (1.5), phrase heads are immediately dominated by phrase nodes (e.g., NP, PP, ADVP, etc.), so modifiers and complements of the phrase head generally appear as phrases or clauses sister to the head. Intermediate levels of structure in the sense of X' theory (N', ADV', etc.) are not represented explicitly in general.
In general the head (N, P, ADV, etc.) is overt and matches the category of the immediately dominating phrase level (NP, PP, ADVP, etc.).
There are cases in which the head is not overt or does not match the category of the phrase level. There are various scenarios in which the category of the word level constituent does not match the category of the phrase that it projects.
There are cases where the head of the phrase has been elided. This can be seen in the lack of a predicate in the IP-ADV-CONJ of (1.9) — a right-node raising construction.
仙台フィルハーモニー管弦楽団 の 第279 回 定期演奏会 は 、 指揮 に ヘンリク・シェーファー 、 ピアノ に 萩原麻未 を 迎え ます 。
A matrix clause is labelled IP-MAT. The pairing of a subject (either overtly expressed in the annotation or not) and a predicate together projects a clause (IP, or ‘inflectional phrase’). Typically, a clause has one main predicate such as a verb (VB), イ-adjective (ADJI), ナ-adjective (ADJN) with copula (AX), or nominal predicate (NP-PRD) with copula (AX). The annotation does not invoke a verb phrase level, and so IP structure is typically flat, with clause level constituents represented at the same level as main predicate material.
花子 が 泣い た 。
この 料理 は おいしい 。
キム は 有能 だ 。
メアリー は 女性 で ある 。
In principle, phrases (rather than words) combine with predicates under IP to compose clauses with complex meanings. In addition to core predicative material (verb, イ-adjective, ナ-adjective + copula, nominal predicate + copula), a small number of other word level constituents can be immediately dominated by IP. These are: coordinating conjunctions (CONJ), single-word interjections (INTJ), auxiliary verbal morphology (VB0, VB2, PASS, AX, etc.), sentential negation (NEG), modals (MD), some formal nouns (FN), and particles (P).
電話番号 を 聞き 違え た か 、 または もう この 電話 は 使わ れ て い ない の だろう 。
When there is a series of one or more phrases which does not form an IP, the construction is a fragment and the member phrase(s) can be placed under FRAG.
まったく 。
= 2011 年 4 月 、 遠野市
おそらく 、 ある 観念 も 。
Single NP or PP elements used as exclamative utterances fall into this scenario.
泥棒 !
This policy applies to imperative utterances formed on こと as well:
この 第6 作 について は 『 仮面ライダー ( スカイライダー ) 』 を 参照 の こと 。
Note that FRAG can appear under CP-THT and CP-QUE:
糞坊主め と はがみ を し た 。
何 の ?
The node label PRN is used for elements bounded by brackets or for an element in an appositive relationship with another element. Dashes and parentheses (PUL and PUR) that set off a PRN will be contained within the PRN node.
Instances of PRN contain a variety of different projections and cover a wide range of various functions some of which are, contrary to their parenthetical nature, semantically significant. In particular, they can be mere side remarks to, functional equivalents of, or even quantifiers of, those which they modify.
○ 国務大臣 ( 三塚博君 )
津波 を 防御 し た 仙台東部道路 の 内陸側 に 立地 する = 仙台市 若林区 荒井東
Among the instances of appositive PRN, the following sorts can be seen:
東北 の 空 の 玄関口 ・ 仙台空港 は 、 16 年 3 月 の 民営化 方針 が 正式決定 し た 。
我々 が ―― 我々 日本帝国人民 が 偉い か 、
答える べき 質問 は -- もう 時間 です ね -- 毎日 の 生活 の 中 で どう すれ ば より 多く の 時間 を フロー の 状態 に できる か
―― それ は 支配人 自身 だっ た 。
そして 私 は その 中 に 現実 の 私 自身 を 見失う の を 楽しん だ 。
社員 全員 が 昇進 を 目指し て いる 。
うち 多賀城市 は 対象 の 農地 97 ヘクタール の 工事 が 全て 完了 し た 。
芸術界 で 今 何 が 起こっ て いる か 動向 が 分かる と さ れる もの でし た
真宗 の 寺院建築 に は 他 に も 内陣 に比べて 外陣 が 広い など 、 他宗 に 見 られ ない 特徴 が ある 。
As a general rule, punctuation is attached at the highest possible level, with the exception that punctuation cannot occur under multi-sentence.
希望者 は 、 住所 、 氏名 、 および 、 電話番号 を 記入 し て 下さい 。
Note especially that when an overt coordinating conjunction follows an IP-ADV, punctuation intervening between the IP-ADV and the following coordinating conjunction is allowed.
彼女 は 一流 の ピアニスト で 、 かつ 優れ た 随筆家 だ 。
As is already mentioned in section 1.3.8, another exception is dashes and parentheses (PUL and PUR) setting off a parenthetical element (PRN). They are contained within the PRN node.
Occasionally texts include metadata such as stage directions or identification of the person speaking. This kind of marginalium is placed under the tag META.
週末 は 人並み が 絶え ない = 昨年 11 月
There is one null element that must be indexed:
Table 1.6: Null element with indexing
| null element | meaning |
|---|---|
| *ICH* | abbreviation mnemonic for ‘interpret constituent here’ |
*ICH* (abbreviating ‘Interpret Constituent Here’) is used as a trace marking discontinuous structures, e.g., rightmost ‘afterthought’ (see section 6.21.4), long-distance scrambling, or other remote relationships that cross phrase levels but are neither calculated (by control, ATB, relative clause extraction), nor specified by sort information (see section 1.5.6 for discussion). Indexing works by adding an index number to the label of the remote (overt) constituent, creating a dummy constituent of the same category in the place where the remote constituent should be interpreted, and adding as its child an *ICH* which shares the same index number. For example, in the following example, a post-posed particle phrase corresponds to an indexed *ICH*. Note that extensions indicating grammatical function (e.g., -SBJ) are marked on the node label dominating *ICH* and not on the node label of the remote constituent.
美しい 街 です よ 、 神戸 は 。
It is possible for a single element to be coindexed with multiple instances of *ICH*, as in examples (1.37) and (1.38):
2012 年 末 の 男性平均寿命 80.18 歳 、 女性平均寿命 84.67 歳 、 戸籍人口平均寿命 82.41 歳 。
訪問診療 の 患者 は 月 20 人 程度 おり 、 定期的 な 巡回診療 は 町外 の 仮設住宅 で 月 3 回 __、 離島 ___の 出島 で 月 2 回 、 江島 で 月 1 回 を それぞれ こなす 。
Note that cases in which constituents may have been reordered leftward but remain at the same clause level (e.g., topicalisation, short-distance scrambling, etc.) are not indicated.
この 服 は 太っ た 人 でも 着 られ ます 。
By contrast a rightmost ‘afterthought’ (see section 6.21.4) is placed at a different level from the clause which should contain a coindexed *ICH*, thereby ensuring the main predicate has rightmost placement:
「 見よ 、 神 の 小羊 」 。
Floating noun-complement clauses adjoining to upper ones also falls under the *ICH*-treatment. The caveat is that these floating clauses are covered by a PRN projection, which it is that undergoes indexicalization. (See also sections 1.3.8, 6.18.3 and 6.17.2)
先生 は また 玄関 の 前 に 戻っ て 、 「 前 へ なら え 。 」 と 号令 を かけ まし た 。
While the single null element of the previous section requires indexing, the null elements of this section are assigned nodes in the annotation scheme without indexing. These unindexed null elements include null expletives, zero pronouns of various kinds, and trace markers of relative clauses. The full list of unindexed null elements used is shown in Table 1.7 along with specifications for their use.
Table 1.7: List of null elements without indexing
| null element | meaning |
|---|---|
| * | zero element |
| *exp* | null expletive |
| *arb* | zero pronoun with generic impersonal reference |
| *pro* | zero pronoun with discourse reference (small pro) |
| *hearer* | zero pronoun with hearer referent |
| *hearer+pro* | zero pronoun with hearer and discourse referent |
| *speaker* | zero pronoun with speaker referent |
| *speaker+hearer* | zero pronoun with speaker and hearer referent |
| *speaker+pro* | zero pronoun with speaker and discourse referent |
| *T* | relative clause trace |
Note that gaps for controlled subject positions (controlees — see section 5.3), and gaps formed by Across the Board extraction (ATB — see section 5.4) are not assigned nodes in the annotation.
A zero element is used when no claims are wished to be made about content but still material is required to support the projection of a constituent. The notable case where this may be required is with the logical subject of a passive clause with (NP-LGS *). This makes no claim about the content of the logical subject besides its need to be present to satisfy the requirements of the passive construction.
財政 は 借金まみれ で 不利益 の 分配 を 迫ら れる 。
Sentences that appear not to have subjects are supplied with a null element subject (NP-SBJ *exp*). This includes:
明日 は 寒く ない だろう 、
非常 に 落胆 し た 石森 だっ た が 、 50 枚 以上 の デザイン画 を 描い た 。
「 火事 だ ! 」
As (1.46) demonstrates, *arb* is used for generic impersonal reference.
論文 を 書く の は とても たいへん です 。
Like an overt pronoun, *pro* may refer to an entity in the discourse context, or may take as an antecedent an element in the same sentence. *hearer*, *hearer+pro*, *speaker*, *speaker+hearer* and *speaker+pro* are more specified versions of *pro*, and are preferred over *pro* when appropriate.
何 を 買っ て あげ よう か 。
___「 先生 、 もう お忘れ です か ? 」
In examples (1.47) and (3.2), zero pronouns mark non-controlled covertly expressed core grammatical roles (NP-SBJ, NP-OB1, NP-OB2), i.e., arguments that are obligatory for interpreting the clausal predicate, but which have null expression and which cannot be reconstructed from either control relations (see section 5.3) or antecedent relations arising from coordination structures, specifically ATB extraction (see section 5.4).
A trace *T* indicates the existence of an empty position in a relative clause modifying the head of a noun phrase that corresponds to the empty position. For more details, see section 3.2.6.1. Example (1.49) includes a trace for a subject (NP-SBJ *T*) to link the head (N 人) with the subject grammatical role inside the relative clause.
わかっ た 人 は 手 を あげ て ごらん なさい 。 」
Note that traces are not limited to core grammatical roles: Traces can indicate the presence of ‘gaps’ in adjunct positions as well as in argument position. In this sense traces contrast with zero pronouns like (NP-SBJ *pro*), which correspond only to arguments with core grammatical roles. For details, about adjunct traces, see section 1.7.3.
An IP-REL that doesn't contain a coordinated clause can have only one local trace. For an IP-REL that contains a coordinated clause, there can be at most one trace per conjunct, although there is no constraint on the grammatical functions of these traces, and no requirement that all the conjuncts contain a trace. For example, in (1.50) the first conjunct in the coordinated IP-REL has a trace in object position, while the second conjunct has a trace in subject position.
彼女 は 皆 が 愛し そして 皆 を 愛する 人 です 。
The second trace in (1.50) is placed to the right of the first conjunct. Otherwise, Across the Board extraction would require co-referential positions with identical grammatical roles in both conjuncts.
One more thing to note about traces: If an overt pronoun or *pro* in a relative clause is co-referent with the trace *T*, this relationship must be indicated by adding sort information (see section 1.8 for details) not on the trace, but on the NP that the modified noun phrase head projects, as seen with example (3.33).
有間皇子 が どう し て も そこ から 逃れる こと の でき なかっ た 悲運
With null elements not being visible, string placement for a null element is often not obvious. Position is therefore determined as follows:
Note that antecedent calculation has an influence on the placement of zero pronouns, as is discussed in the remainder of this section.
よそ から 、 もらっ た お酒 が 二 升 あっ た 。
この ビル は 、 現在 の 耐震基準 に 照らし合わせる と 、 強度 が 不足 し て いる .
The calculation of antecedents for empty positions can affect the placement of zero pronouns. For example, the calculation observes a rule for co-reference into coordinate clauses called ‘Across the Board’ extraction (ATB). Specifically, as Figure 1.1 illustrates, any overt category x in IPn must be an antecedent to one empty position (preserving its grammatical role) in each of coordinate clauses IP1, 2...n-1 disambiguated with CONJ provided that x is sister to and precedes IPn-1, and provided that IPn-2 is the first element in IPn-1, and provided that IPn-3 is the first element in IPn-2, and so on. This requires that each of IP1, 2...n-1 ‘inherit’ the antecedent relation from the category in IPn. Placement of a zero pronoun in IPn may be decided by whether that zero pronoun is an antecedent for all non-final conjuncts (in which case the zero pronoun is leftmost).
Figure 1.1: ‘Across the Board’ extraction (ATB).
「 和国 の 教主 」 として 尊敬 し 、 観音菩薩 の 化身 として 崇拝 し た 。
In (1.55), the zero pronoun subject of the final clause (the matrix clause) is the antecedent of an empty position in the non-final clause (the precedence of the coordinated clauses). In the non-final clause a null pronoun for a primary object is placed in initial position, while the final clause contain a different overt primary object.
火鉢 に かざし て 、 文字 を あぶり出し た .
In (1.56) as well, the zero pronoun subject of the final clause (the main clause) is the antecedent of an empty position in the non-final clause (the precedent of the coordinated clauses), but the zero-pronoun primary object is not. The non-final clause has its own overt primary object. In this case, the zero-pronoun primary object of the final clause must be placed in non-initial position, in violation of the default, after the non-final clause.
「 私たち の 話 を 聞い て 、 監視 し て いる の よ 」
Following the default practice of placing zero pronouns in clause-initial position could establish an undesired control relation (see section 5.3 for details). Such a situation is caused by the antecedent accessibility hierarchy, a part of the calculation of antecedents for co-reference relations which stipulates that controlled positions are more accessible by object-antecedents than by subject-antecedents. Given the antecedent accessibility hierarchy, subject-antecedent control into a subordinate clause requires that no object appear to the left of that clause. See an example of this scenario below. A clause IP2 headed by a transitive verb but not containing an overt NP-OB1 argument, shares a subject with a subordinate clause IP1 , but does not share an object. The more accessible NP-OB1 argument cannot appear to the left of IP1, lest it bind into IP1 (i.e., the former be interpreted as the subject of the latter), so it is placed after IP1.
急い で 食べ ない で 。
For other configurations, there are fewer conditions on drawing co-reference relations. The clause marked with a conjunctive particle たら in (1.58) contains an overt object PP that co-refers to the null subject of the main clause. Precedence is a sufficient condition for inferring the possibility of co-reference here, and this motivates the non-initial placement of *pro*. Placing the zero pronoun (NP-SBJ *pro*) of the matrix to the right of the IP-ADV conditional clause allows the PP-OB1 argument of the preceding IP-ADV to be taken as antecedent for that zero pronoun, although without some extra information, co-reference is only left open as a possibility, rather than being completely determined. In example (1.58), the PP-OB1 それを of predicate あげ(る) in the subordinate clause and the null pronoun subject of なくなった in the matrix clause (NP-SBJ *pro*) uniquely share ‘sort information’.
それ を 三 つ 彼 に あげ たら なくなっ た 。
In (5.68), a subordinate clause and a main clause share both a null subject and a null object. In such a scenario, a null pronoun object cannot be placed at the begining of the main clause, lest it control into the subject position of the subordinate clause. Accordingly, null pronouns sharing ‘sort information’ are introduced into both clauses, with the object of the main clause being placed after the subordinate clause.
ダウンロード し て 印刷 すれ ば 、 学校 や 家庭 で 手軽 に 取り組める 。
This section explains how core grammatical roles are annotated in different environments. Section 1.7 will describe how peripheral grammatical roles are annotated.
First, we address overt instances of NP with core grammatical roles in the simple clause. Core grammatical roles are the roles that nominal arguments take with respect to the predicates that select them. An argument is, in principle, a constituent without which the predicate cannot be fully interpreted. However, there are many instances where it is hard to judge whether a constitutent that co-occurs with a given predicate is obligatory or optional. For example, it is difficult to judge the status of path denoting elements co-ocurring with predicates of motion. As another example, the status of quotational phrases with regard to verbs of cognition and communication may be ambiguous. At present our policy with regard to the first scenario is to mark paths as primary objects when they appear overtly, but to not require null element paths to be annotated for verbs of motion otherwise. Our policy with regard to the second scenario is to treat quotational phrases as having the same status (CP-THT-OB1) with respect to a verb like 言う as to a verb like 話す, but to indicate their adverbial status (CP-THT-ADV) when appearing with verbs that don't involve propositional content at all (see section 6.18.5).
Core grammatical roles are identified by node label extension -SBJ (subject), -SBJ2 (secondary subject), -LGS (logical subject), -OB1 (primary object), -OB2 (secondary object), -CMPL (complement), and -CZZ (causee). They are typically marked by particles (P-ROLE) が, に, を, の (see section 2.3 for details), etc. However, they may also be marked by toritate particles (P-OPTR) like は and も. They can also appear unaccompanied by particles. This means that the grammatical role of a constituent cannot be reliably predicted solely on the basis of the form of the attaching particle.
When a constituent has a core grammatical role, an extension indicating that role is added to the node label. In (1.60), the particle phrase arguments (PP 漁夫が), (PP その女房に) (PP 金を) are marked -SBJ (subject), -OB2 (secondary object), and -OB1 (primary object), respectively.
漁夫 が その 女房 に 金 を 渡し て いる ところ だっ た 。
We assume that the realisation of a constituent with the grammatical role of subject is obligatory for virtually every predicate that heads an IP, so that 1-place predicates have only the subject as argument. Subjects are generally able to serve as antecedents to reflexive 自分 and are able to license honorification morphology on the predicate. Semantically subjects are typically Agents of transitive and unergative predicates, Patients of unaccusative predicates, and Experiencer or Possessor or Attribute-holder of stative predicates.
As a general principle (noted in section 1.3.6), every IP has a subject, and every subject/predicate pair projects an IP. Sentences that appear not to have subjects are supplied with a null element subject (NP-SBJ *exp*) (see section 1.5). The principle that all IP layers have subjects (either expressed by phrasal categories or determined by control) is not intended to be a claim about Japanese clause structure in general. Requiring fully saturated argument structures simplifies the model and serves as one way to check accuracy of annotation. Using (NP-SBJ *exp*) involves the judgments of native speaking annotators, rendering ‘subjectless sentences’ easily searchable and accessible for evaluation, given that (NP-SBJ *exp*) distingushes them from controllees. Accordingly, an IP that is not a potential control environment (viz. IP-MAT, IP-SUB, IP-REL) always contains a subject. An IP that is potential control environment (viz. IP-ADV, IP-SMC, IP-EMB) does not contain an NP-SBJ if an immediate upstairs argument controls into that IP. See sections 5.3 and 5.4 for examples of the implementation.
For arguments other than subjects, we refrain from associating their grammatical roles with traditional semantic descriptions. Note that a non-subject argument in a 2-place predicate is formally an -OB1, i.e., the primary object in the sentence. Thus for example a non-subject に-marked argument that might not be regarded as ‘direct object’ in traditional grammar will nevertheless be assigned the label extension -OB1.
散歩 を し て いる うち に 、 来る とも なく 駅 まで 来 て しまっ た 。
If an argument appearing with a 3-place predicate is (or can be) marked by the particle を, we regularly mark that argument as -OB1. The に-marked beneficiary NP of donative verbs is assigned the grammatical role -OB2.
叔父 は 花子 に 小遣い を 与え た 。
In the ‘double subject construction’, the second subject (viz. the one closest to the predicate) is frequently marked by が. The constitutent gets the extension -SBJ2 on its node label.
太郎 は 背 が 高い 。
Note: It is not the case that every element marked as -SBJ is actually a subject of the local predicate. In a context where -SBJ precedes and is sister to -SBJ2, in fact -SBJ2 is interpreted as the subject of the predicate that heads the containing clause. An exhaustive search of subjects expressed with nodes would thus require excluding -SBJ when it is sister to -SBJ2, and including -SBJ2 in its place.
In a passive sentence, に and some other particles mark the semantic Agent of the action denoted by the core predicate, i.e., the ‘logical subject’. In this case, the grammatical role is indicated by the label extension -LGS.
ジョン は 先生 に しから れ た 。
NP instances which perform a core grammatical role may not be necessarily marked by a particle associating that role with a predicate. To begin with, they may occur with toritate particles like は and も. In these cases as well, an extension is added to indicate the core grammatical role. In (1.65), the PP headed by は receives the extension -SBJ.
花子 は まだ 学生 に 見える 。
When a particle is omitted, there is no reference to the omitted particle and the NP is linked directly under a label specifying the grammatical role, e.g., NP-SBJ, NP-OB1, etc.
君 、 あの 本 読ん だ 。
There exist predicates selecting arguments which don't have the phrasal category NP or PP, most importantly CP-THT and IP-SMC. They are distinguished from adjuncts (see section 1.7 on adjuncts) by extension tags (-OB1, -SBJ etc.). For detais, see section 6.18 on CP-THT and sections 5.1.4, 4.6, 6.10 and 6.12 on IP-SMC.
Furthermore, in constructions with verbs する and なる preceded by onomatopoeic or demonstrative adverbs, the entire adverbial phrases serving as complements are marked up as ADVP-CMPL.
ナシャ が イライラ し て 言っ た 。
こう なっ た 以上 、 もう どう し よう も ない 。
When an argument with a core grammatical role does not appear explicitly in the sentence, it may be a controllee controlled by an NP in a higher clause, or a gap corresponding to an ATB extracted element. In these cases the ‘gap’ is left empty, without any node. In the case of a relative clause, the ‘gap’ is marked up with a trace as a terminal node: *T*. All other unexpressed arguments are treated as zero pronouns, with terminal nodes in the tree chosen from a closed set: e.g., *pro*, *exp*, *arb*, *hearer*, *speaker*, *hearer+speaker*. These are put directly below a phrasal node with a label showing the grammatical role like NP-SBJ, NP-OB1, NP-OB2, NP-SBJ2, NP-LGS. For details, see section 1.5.
けれども 、 いつ 来る か 、 わから ない 。
When a subject NP is omitted in a subordinate clause, its reference may be determined by a controller in the matrix clause. In such cases, no corresponding node is added to our annotation. In (1.70), the reference of the subject of the verb embedded by で is provided by the controller, i.e., the subject of the matrix clause わたし.
急い で わたし は 出かけ まし た 。
Peripheral grammatical roles are performed by adjuncts (adverbial clauses, adverbial phrases, and various kinds of PP). Adjuncts are not essential to the interpretation of the predicate: They are optional constituents.
Nominal adjuncts marked by particles に, へ, で, から, まで, と, etc. form PP adjuncts. For PP adjuncts, the node labels are desired to have the extensions depending on their semantics such as -LOC (spatial location), -TMP (temporal location), -MSR (temporal extension or frequency), or -ADV (other adverbial function), as the following examples show.
あそこ の ベンチ で 食べ ましょ う 。
つぎ の 汽車 は 七 時 に 出る 。
この 手 の 話 は 何 回 も 聞き まし た
太郎 は 500 メートル も 泳げ た 。
However, unlike PP arguments with core grammatical roles, PP adjuncts do not always have node label extension in the current annotation. There are many PP instances without extensiton, like PP instances projected by the particles で and に in (1.75).
ビル は 列車 で パリ に 行っ た 。
There are also particles like に that can mark either an argument or an adjunct. With regard to the obligatory usage of に, see section 1.6.1.
When NP or PP instances formed without role particles function as adjuncts, depending on their semantics, their node labels can receive the extensions -LOC (spatial location), -TMP (temporal location), -MSR (temporal extension or frequency), or -ADV (other adverbial function).
In the examples below, the expressions あちこち and この町は denote a spatial location.
両親 は 子供 を ___あちこち 連れ回し た .
この 町 は いい レストラン が 少ない です 。
In the examples below, the expression 先週の土曜日 and きのうは indicate a point in time at which an event occurred.
先週 の 土曜日 、 街 で 昔 の 友人 に 会っ た 。
きのう は かぜ を ひき まし た 。
In the examples below, the expressions 5年間 and 食事中は denote the duration (extension) of an event or state. In the following example, the expression 一度 denotes the frequency (distribution) of an event.
また 5 年間 使い 続ける ぞい !
食事中 は テレビ を 見 ませ ん 。
私 は 、 途中 で 一 度 、 悪い 夢 を 見 た 。
For negations and other stative predicates, disambiguation between -TMP and -MSR may seem more difficult at times, but it is useful to remember that the end of the duration denoted by -MSR implies a change of state or end of event, whereas this usually is not the case with -TMP.
今日 は 、 誰 に も 会わ ない 。
If a combination of expressions ‘measure out’ an event, the expressions are treated as instances of NP-MSR:
In the examples below, the expression 予定通り and 本当は have an adverbial function which does not correspond to any of the more specific categories.
会議 は 、 予定通り 3 時 に 始まっ た 。
本当 は 叔母 の アイリーン です が
If a particle is omitted from an NP adjunct, a PP projection is created and the expected particle is inserted as a terminal node that is placed between instances of ‘*’.
それ 、 おれ 行く ん だ けど さー
See also section 2.14.
Aside from an ATB environment a ‘gap’ in a relative clause is the only place where an unindexed null element can be associated with an adjunct role. If the adjunct role is temporal or locative, this information can be added as an extension of the label. For example, (1.87) has trace: (NP-LOC *T*).
ここ が 高津さん が 講演 し た ところ だ 。
If the adjunct role associated with the modified head N is neither temporal nor locative and usually indicated by a particle, the trace can be located in an NP under a PP with the missing particle specified in a terminal beginning and ending with ‘*’. For example, (1.88) has trace: (PP (NP *T*) (P *に*)).
早朝 、 彼 は 住み慣れ た 町 を 出で立っ た .
Sort information is used for resolving reference for pronouns. Pronouns such as これ、あいつ、そこ, etc., and the null pronoun *pro* trigger the need to introduce sort information that establishes a link with the accessible antecedent.
If an NP X (either nominal or pronominal) both c-commands and shares sort information with a pronoun Y, and there is no closer candidate to Y which bears the same sort information, an antecedent relation will be calculated between X and Y.
If a nominal NP X precedes pronoun Y in the discourse, and shares sort information with Y, and there is no closer candidate to Y which bears the same sort information, an antecedent relation will be calculated between X and Y regardless of their respective structural positions.
One exception is when an indefinite element is inside of an ‘operation of closure’ such as the scope of negation or modality. In such environments, only definite elements can serve as antecedents to outside pronouns.
Note that an identity relation obtaining between two nominal expressions will not be calculated through sort information. Such information can be safely added, if desired, and such information is potentially useful for linguistic studies and a full representation of the semantics of texts, but it is outside the purview of Scope Control Theory. Cataphoric relations are also exceptions in this sense, but are interesting and worth marking up.
This section details the situations where sort information is needed.
しかし 、 その 内容 は 『 タイガーマスク 』 に 言及 し 、 「 自分 も 仮面 を 被れ ば ヒーロー に なれる 」 という 児童 の 願望 を 指摘 し て いる こと 、 主人公・九条剛 が 普通 の 体育教師 で 鍛錬 によって ヒーロー の 力 を 得 て いる など 、 当時 流行 し て い た スポーツ根性もの の 影響 が 強い 内容 で あっ た 。
While IP-SMC under CP-THT allows control from an immediately higher IP, IP-SUB under CP-THT is an island for control. In (1.90) below a c-commanding pronoun becomes an antecedent for another pronoun in a quotation by virtue of sort information.
彼 は たびたび 僕 に 「 私 は もう 長く は ない 」 と 言い 、 僕 は それ を 戯言 と 思っ て い た 。
In (1.91) a trace *T* shares sort information with a null element *pro* in a quotation via the information on the antecedent of the trace (namely, the whole NP).
そして もう 1 つ 欲しい と 思っ た の は もっと 精巧 な 技巧 と 手法 でし た
In (1.92) below there is an extra NP layer over IP-EMB, preventing control from an upstairs subject argument. Sort information is added to the IP-EMB to establish both the relationship between the subjects, as well as an identity relationship between the object in the IP-EMB and the NP taking the modified head in a complement PP.
「 確か に 私たち は できる 限り の こと を し た わ 、
Cataphoric relations (where an pronoun precedes a co-valued element) are rare, and typically obtain across sentence boundaries. These don't actually get represented in the visualisation. The pair of examples in (1.93) and (1.94) show a null pronoun co-valued with a following quotation:
叔母 は イライザ が ため息 を つく まで 待ち 、 それから 言っ た 。
「 ああ 、 それでは 、 あの 方 は より よい 世界 へ 行っ て しまっ た の ね 。 」
Here is another pair of examples with an overt pronoun covalued with a following nominal element:
あれ が 一 つ 手に入っ たら ね 、
最新式 の 馬車 で オルーク神父 が 言う に は 騒音 が ない ん です って 、
The file aozora_Natsume-1908.psd also starts with a cataphoric reference to the following discourse. It is not possible to mark up something like this in our present framework.
These are rare, typically appearing in noun modifying clauses. In such cases, the practice is to associate the pronoun in the modifying expression with the nominal head using sort information.
Furthermore, the sort information is supported by the use of a trace, complement of a dummy particle.
(PP (NP *T*) (P-ROLE ***))The function of (P-ROLE ***) is to push the reference of the modified NP into the domain of discourse so that it can serve as the antecedent for the pronoun in the modifying clause.
有間皇子 が どう し て も そこ から 逃れる こと の でき なかっ た 悲運
この, その, あの function both as possessive pronouns and determiners. This is a holdover from when こ and そ were pronouns. In their use as possessive pronouns, the relevant sort information is needed. Their distribution requires that sort information be added to the part of speech tag D. The possessive pronoun function is most easily discernable when these words are attached to relational nouns: その他, その後, その残り, etc. But other examples can be found:
義姫 が 実家 の 山形城 へ 突如 出奔 し た の は この 4 年 後 で ある こと が 一 次 史料 から すでに 明らか に なっ て いる 。
家康晩年 の 1616 年 ( 元和 2 年 ) 1 月 23 日 の イギリス商館長 ・ リチャード・コックス の 日記 で は 、 「 風評 によれば 、 戦争 は 今や 皇帝 ( 家康 ) と その 子 カルサ様 ( 松平上総介忠輝 ) と の 間 で 起こら ん と し 、 義父 政宗殿 は 、 カルサ殿 の 後援 を なす べし 云々 」 と 記さ れ て いる 。
Capturing a partitive relation between a topicalized superset and a quantified referring expression is done with the use of a ‘dummy’ particle ((P-ROLE ***)) on the topic NP and a ‘dummy’ genitive postpositional phrase (PP (NP *pro*) (P *の*)) modifying the referring expression:
自分たち は ほとんど が 次男坊 ばかり です から 」
In some texts pronouns appear without overt antecedents. In order to establish anaphoric relations, then, dummy discourse antecedents must be set up.
These provide antecedents for subsequent pronouns:
今度 は 彼 に 望み は なかっ た 。
Depending on the structures involved, the employment of sort information can open up multiple possibilities for establishing dependencies. For example, there are at least two alternatives for marking the sentence 「関連産業の集積や教育レベルの向上など波及効果は計り知れず,東日本大震災からの復興にも貢献できると期待される。」 so that the necessary dependencies are calculated. Both require sort information, and the truth conditions are the same for both strategies, but the information structure is subtly different. In (1.103) the は-marked element is under the scope of negation, and the clauses are coordinated. In (1.104) the は-marked element is in a matrix subject position and the first predicate heads a subordinate clause.
関連産業 の 集積 や 教育レベル の 向上 など 波及効果 は 計り知れ ず 、 東日本大震災 から の 復興 に も 貢献 できる と 期待 さ れる 。
関連産業 の 集積 や 教育レベル の 向上 など 波及効果 は 計り知れ ず 、 東日本大震災 から の 復興 に も 貢献 できる と 期待 さ れる 。
Null pronouns that receive a plural interpretation by virtue of having multiple antecedents A and B are rare. The mark-up practice is to add sort information to each of the antecedents (e.g., (NP;{aaa} (N A)); (NP;{bbb} (N B)), etc.), to concatenate these with +'s in the sort information of the pronoun in question (e.g., (NP;{aaa+bbb} *pro*)), and to introduce a preceding dummy constituent of the form (immediately to the left of the pronoun).
(ANT;{aaa+bbb} *)仙台市 に 拠点 を 置く チーム の 担当記者 が 「 見 た スポーツ の 力 」 、 福島県 から 宮城 、 岩手 両県 の 沿岸支局 に 転勤 し た 記者 が 「 福島 と 津波被災地 と 」 を テーマ に 思い を つづっ た 。
Plural antecedents (e.g., (NP;{TWINS} 双子)) can be related to subset-denoting null elements by introducing two immediately adjacent dummy constituents, the first of which consists of a *pro* that shares sort information with a plural antecedent of the form:
(PP (NP;{TWINS} *pro*)and the second is an operation that splits the reference:
(PP (NP (CONJP (NP;{JOHN} (D *)))The two dummy constituents together serve to relate the antecedent (NP;{TWINS} 双子) to a null pronoun such as (NP;{JOHN} *pro*).
The following example shows how this technique for splitting an antecedent can be applied to referring expressions in an internally headed relative clause:
警官 が 泥棒 を 追いかけ て い た の が 二人 共 川 に 落ち た 。
Bridging inferences occur when a relationship not expressed overtly is assumed between two discourse referents (Asher and Lascarides 2003). For example, a natural way to interpret the discourse 「太郎はあるレストランに行った。ウェイターの態度に不満を抱いた」 (‘Taro went to (a) restaurant. He was displeased by (the) manners of (the) waiters.’) is to assume that the definiteness of ‘the waiters’ depends on a (covert, assumed) belonging relationship with respect to ‘restaurant’. There is an annotation convention that can be applied when such a relationship has consequences for the information state of one of the said expressions, as with the ‘the waiters’.
The two examples below form a discourse in which the reference of ‘代表的な例’ is dependent on the reference of ‘童話など’ in a preceding sentence. A dummy PP headed by *の* and containing an NP complement headed by <tt>*pro* is linked to the preceding expression with sort information:
童話 など に も 数多く 登場 し 、 多く は 主人公 に 倒さ れる 。
代表的 な 例 として は 、 一寸法師 、 桃太郎 など が あげ られる 。
This chapter details the annotation adopted to describe the behaviour of particles (also known as postpositions), which are tagged P together with a node label extension (-ROLE, -CONN, -COMP, -FINAL, or -OPTR). Particles will typically attach to a noun phrase or clause and carry out various functions. In terms of consequences for the annotation, particles can be grouped into the following types:
Note: Particles that share a given phonological form can mark more than one function or role in the same group. For example, in the group of particles for core grammatical roles, に can mark -SBJ, -SBJ2, -LGS, -OB1, and -OB2. Furthermore, particles that share a given phonological form can appear in more than one group. For example, と appears in the group of particles for core grammatical roles (P-ROLE), the group of particles for peripheral grammatical roles (P-ROLE), the group of conditional conjunctive particles (P-CONN), and the group of complementizer particles (P-COMP). Conversely, a particle of a given group and function may have more than one phonological form.
The term ‘case particle’ is defined here as subsuming the class of particles for core grammatical roles and the class of particles for peripheral grammatical roles, while excluding all other classes. In the annotation here, these classes share the P-ROLE node label.
When a sequence of words as a whole functions like a single P, it is chunked as an single complex particle and labelled P. Many of these complex particles are chunked in accordance with the LUW (Long Unit Word) analysis in BCCWJ and CSJ. Of course, for many such sequences, there may be cases where one or more components are functioning as lexical items. Such cases require re-segmenting by hand with each segment reassigned to its appropriate structural position.
There are over 60 complex particles represented here, a rather long list for a functional category. Here are examples of some complex particles:
For example, the complex particle によって can be associated with the core grammatical role -LGS in passivized predications, as in (2.1).
この デザイン画 は 平山 によって 渡邊専務 に もたらさ れ た 。
The complex particle について in (2.2) marks an adjunct.
モーセ は 、 わたし について 書い た の で ある 。
However, について in (2.3) is not analyzed as a single particle, but as a sequence ending with a verb.
彼 は 新しい 仕事 に つい て 、 まるで 生き返っ た よう だ .
The contrast above is trivial as the sense of the word つく differs greatly between the examples.
The contrast between the complex particle につれて in (2.4) and the particle-verb sequence in (2.5) below is also easy to make because the complex particle almost always follows verb forms.
経済 が 発展 する につれて 、 社会 の 矛盾 も 拡大 し て き た 。
ディズニーランド に 連れ て 行か れ まし た 。
But for a complex particle like にむけて/にむけた more difficult distinctions must be made. In (2.6) below, the question is whether 向け takes a subject, an object, and a goal argument, or not. As it does not have a full argument structure here, we treat it as the complex particle にむけて/にむけた.
「 国際交渉 に向け 日本政府 の 意思表示 が 必要 だ 」 と 国 が 早期 に 誘致 を 決断 する よう 訴え た 。
Particles for core grammatical roles head instances of PP with core grammatical roles. However, there is no one-to-one correspondence between role and particle. Consequently, PP nodes with core grammatical roles are given node label extensions (PP-SBJ, PP-OB1, PP-OB2, PP-CMPL, PP-SBJ2, PP-LGS, PP-CZZ). Thus, for example, a subject PP headed by P-ROLE が will receive the node label PP-SBJ.
An example with P-ROLE instances marking NP nodes with various roles: が with NP-SBJ, を with NP-OB1, and に with NP-OB2:
それ が 私 に 何とやら 奇妙 な 感じ を 与え た の で ある 。
Role particle が frequently marks a subject (-SBJ). However, other role particles like に can also mark subjects (see 2.3.3,2.3.9). In particular, in noun modifying clauses particle の can also mark subjects (see section 2.3.5)
Role particle が marks not only subjects, but also other grammatical roles. It marks primary objects (-OB1) with the following kinds of predicates. Note, however, that the same grammatical function can often be expressed by a particle other than が for these same predicates. For example, the primary object of a potential predicate can be accompanied by を as well as by が.
太郎 は 財産 が ある 。
スミスさん は 中国語 が 分かる 。
救急車 が 必要 です か 。
私 は 虫 が 嫌い です 。
おしぼり が 欲しい の です 。
Role particle が can also mark a secondary subject (-SBJ2) in a ‘double subject’ sentence:
象 は 鼻 が 長い 。
Role particle を can mark the core argument role of primary object (-OB1). In the annotation here, this includes not only canonical ‘direct objects’, but also the ‘source’ arguments of verbs denoting departure or separation such as 出る, 降りる, 発つ, 出発する, 離れる, やめる, 卒業する, ‘path’ arguments of verbs denoting motion such as 歩く, 走る, ジョギングする, 旅行する, 飛ぶ, 通る, 渡る, 行く, 来る, ‘time period’ arguments of verbs denoting passage of time such as 過ごす, 明かす, 暮らす.
午前 6 時 に 家 を 出 まし た 。
私たち は ショッピングセンター を ぶらぶら 歩い た 。
私たち は トランプ を し ながら 楽しい 時間 を 過ごし た 。
However, when a verb of motion does not co-occur with a path element, it is not necessary to add an object null pronoun corresponding to that unless the context suppies some independent reason.
ビル は ベルリン に 車 で 行っ た 。
With regard to verbs denoting passage of time, expressions denoting periods of time may co-occur with these unaccompanied by any particles. In the absence of some positive reason to mark such expressions as objects, they are annotated as NP-MSR. For details on the use of the extension -MSR, see section 1.7.2.
何 を する か と 言う と 誰 か の 家 で 1 週間 過ごし ます
It is rare for role particle を to appear more than once in the same clause. See section 6.8 for details on this type of construction.
Role particle に marks the secondary object (-OB2) of many donatory/benefactive and other 3-place verbs, such as やる, あげる, もらう, くれる, 与える, 貸す, 借りる, 渡す, 届ける, 贈る, 提出する, 聞く, 教える, 伝える, たずねる, 報告する, 命じる, 相談する, 紹介する, etc. For example, see (2.7). In the following example, the secondary object is marked with に, while the primary object is indicated with a null pronoun *pro*.
そこで 彼ら は 、 もう 一 度 この 盲人 に 聞い た 、
Role particle に marks the subject of predicates denoting possession (e.g., ある, いる, ない, etc.), potential and cognition (e.g., できる, 分かる, 見える, 聞こえる, etc.), necessity (e.g., 必要だ, 不要だ, 要る, etc.), and feelings (e.g., うれしい, 悲しい, etc.). These predicates overlap with those that mark primary objects with が (see section 2.3.1).
実 を いう と 、 私 に も その 理由 は わから ない 。
In principle, the non-subject argument in a 2-place predicate is assigned the core grammatical role -OB1. Accordingly, there are predicates for which the primary object is marked by role particle に. Below are a few typical sorts of scenarios.
その 手 に さわっ た 革財布 。
犬 が 飼い主 に 甘える
私 は ハタと ある 事 に 気がつい た の です 。
彼女 は 経営者 に ふさわしい
Constitutents headed by に that contain agents in (て)もらう and (て)ほしい constructions, etc. are also tagged as -DOB1 (derived primary object). In such constructions, a small clause (IP-SMC-OB1) preceding the predicate is a control environment inheriting the reference of a preceding に-marked derived object (-DOB1). For the (て) ほしい construction, see section 4.8.
文章 に する とき は さ 、 ワトスン 、 兎 に 働い て もらう ん だ ね 」
みなさん に は 、 限りある 時間 で 助け 合い 、 命 を 守る すべ を さまざま な 視点 で 考え て ほしい 。
に can also be used to mark the logical subject of passive sentences. In such scenarios, the phrase に projects is tagged as PP-LGS.
ジョン は 先生 に しから れ た 。
When PP instances headed by に serve as arguments of verbs する and なる and cannot be regarded as being predicated of another argument, they are given the extension tag -CMPL (for cases where predication holds, see section 4.6).
日本人 が 働き者 だ という こと を 我々 は よく 耳 に する 。
- 過去 の 大災害 で の 支援活動 を通して 、 参考 に なる 事例 は 。
Role particle に can also mark adjunct (optional) elements (see section 2.4). However, the distinction between core and role (that is, the function of に marking -OB1/-OB2 as opposed to the function of に marking adjuncts) is sometimes difficult to make. Furthermore, some predicates (for example, 当たる) are used in a variety of senses, some of which select a に-marked -OB1 (e.g., 役目に当たった) and some of which do not (e.g., 宝くじが当たった).
Role particle と marks primary objects of predicates such as 似る, 会う, 喧嘩する, デートする, いちゃつく, etc.
彼女 と は 、 去年 会っ た きり 、 手紙 も 出し て い ない 。
だって 彼女 は いつも 風 と いちゃつい てる ん だ から 」
In many instances these predicates are reciprocal predicates in which a plural subject element can saturate the argument structure:
ヘプバーン は 「 私 と ユベール は よく 似 て い ます 。 好み が 同じ な の です 」 と 語っ て いる 。
And also と marks secondary objects of a variety of 3-place reciprocal predicates.
彼 は 、 弟 と 一 等 賞 を 競り 合っ た .
When a PP headed by と serves as an argument of verbs such as する, なる, etc. and cannot be regarded as being predicated of another argument, it is given the extension tag -CMPL (for cases where predication holds, see section 4.6).
英雄 を 襲う の は 、 黒々 と し た 死 の 闇 で ある 。
一体 支那人 は いざ と なる と 、 覚悟 が 好い 。
The particle と in naming constructions with verbs such as 名付ける or 呼ぶ is treated as P-COMP (see section 2.7.3).
Subjects are typically marked by が within noun-modifying clauses, but の can also mark a subject in this context.
これ は 、 わが 社 の 開発 し た 最新型 の ワープロ です 。
私 の 知ら ぬ お名前 で あっ た 。
As previously mentioned in section 2.3.1, role particle が can mark the primary object of some predicates. In noun-modifying clauses, role-paticle の can mark primary objects in place of が.
高価 な 物品 を 購う こと の でき ない 者たち は 、 草鞋 を 作っ て いる 。
The logical subject of a direct passive construction is typically marked by role particle に (see (2.27) for an example). In the corpus are also found examples of such logical subjects marked with role particles によって, により, から, より, and で.
東大寺 は 745 年 に 聖武天皇 によって 建て られ た 。
本山 は 、 足利義満 により 建立 さ れ た 京都 の 相国寺 。
ワシントン の 桜 は 、 1912 年 に 日本 から 贈ら れ まし た 。
質疑 を 終わり まし た ところ 、 日本共産党 を 代表 し て 西山委員 より 修正案 が 提出 さ れ まし た 。
ああ 、 この 土地 の 南 も 北 も 、 皆 敵 で 埋め 尽くさ れ て いる
The causee argument of a causative construction is typically marked by role particle に or を (see (2.44) and (2.45) for examples.) In the corpus are also found examples of such causee arguments marked with role particles から, において, に対し, に対して, について, へ, and をして.
これ は 神 の 子たち が 人 の 娘たち の ところ に はいっ て 、 娘たち に 産ま せ た もの で ある 。
わたし は 、 その 人々 を 終り の 日 に よみがえら せる であろう 。
独裁者 が 部族 に対し その 降伏条件 に 無理矢理 同意 さ せ た 。
しかるに 女子解放運動 は 、 女子 を して その 母性 を 失わ しめる から 宜《よろし》く ない 。
For many 3-place predicates, including donatory/benefactive verbs, the secondary object (i.e., the third argument) is marked by role particle に (see section 2.3.3). In this corpus we also find role particle から and (complex) role particles に対し and に対して used to mark secondary objects.
その 家 は 我々 から 光 を 奪っ た 。
なお 、 本件 に対し 附帯決議 が 付さ れ て おり ます 。
Particles other than typical core particles like が or を can mark subjects and primary objects. Lists of role particles that mark these arguments, respectively, are presented below:
In the following examples, から and で are used to mark the subject role.
鈴木さん に は あなた から 伝え て 下さい 。
後 は 私達 で やり ます 。
In the following examples, で and について are used to mark the primary object role.
彼 の 人生 は 困難 で いっぱい です 。
私 は その 問題 について じっくり 考え た 。
Deciding what is optional depends not only on the particle but on its relationship with the predicate in question. An allative particle such as まで (or に in some contexts) may be considered to mark optional elements for a basic verb of motion such as 行く, but for motion verbs like 到着する and 来る, an endpoint could be considered an essential part of the interpretation, leading an annotator to mark up まで in an expression like 在来線がとうとう村まで来た as projecting PP-OB1. Again, a role particle へ may be optional for a basic verb of motion such as 行く (as in (2.56) below, for example), but might be considered an argument when relating to a predicate like 向かう. A list setting out the argument structures of common verbs would resolve many of these details, but in the absence of this, annotators have to proceed with some sensitivity to these issues.
While role particles may themselves reveal information about grammatical role, the annotation can still benefit from further specification, especially in the case of extremely polysemic particles such as に. For example, in (2.54) below, there are two particle phrases headed by に, one of them temporal location, and another marking spatial location. These functions can be specified, respectively, by the extensions -TMP and -LOC.
1997 年 に は 仙台市泉区 に ホームスタジアム で ある 仙台スタジアム が 建設 さ れ た 。
In addition to temporal and spatial location as exemplified above, role particle に can mark goals of motion (山に登る), states resulting from change (日本語に翻訳する), purposes of motion (釣りに出かける), causes (酒に酔う), bases for frequency or proportion (一日に三度ご飯を食べる), etc. (Masuoka, and Takubo 1992, Nihongo 2009). Among these, the extension -PRP is attached to に-headed PP instances denoting purposes of motion, as in the example below. For details on annotation practice in instances where に attaches to clauses denoting purposes of motion, see section 6.16.
彼女 の お母さん は 買い物 に 行き まし た 。
Adding the functional information presented in the examples above can only be done by annotators' intuitions and is time-consuming. Therefore, there exist PP instances that do not have functional information labels in the current annotation.
次 の 日 、 ツバメ は 波止場 へ 行き まし た 。
爺さん は 酒 の 加減 で なかなか 赤く なっ て いる 。
ITEL は 1988 年 から 1992 年 まで APCOM を 所有 し て い た 。
私 は アンディー と 一緒 に パーティー に 行き まし た 。
Note again that a ‘partner in action’ may need to be specified as an argument (-OB1 or -OB2), depending on the predicate it appears with (see section 2.3.4 for details).
Role particle の is first and foremost a genitive marker marking a relationship (possessor::possessee, whole::part, set::member, etc.) between one noun phrase and the head of another (see 2.4).
私 の 心 は 、 不安 と 期待 に 掻き乱れ た .
日本語 の 勉強 の 時間
When a head noun denotes an action, a の-headed complement PP can denote elements that have the same semantic roles as arguments. For example the の-headed complement PP in (2.62) has the semantic role of theme with respect to 決定 and the の-headed complement PP in (2.63) has the semantic role of agent with respect to 決定. However, as these are not arguments of predicates, they aren't supplied with grammatical role information.
どれ を 優先 する か の 決定 は 、 今後 の 課題 だ 。
4 内閣総理大臣 は 、 ___前項 の 規定 による __閣議 の 決定 が あっ た とき は 、 遅滞 なく 、 基本方針 を 公表 し なけれ ば なら ない 。
Note that the form の appears as the adnominal form of the copula, so care is needed to distinguish this from the genitive particle (see section 4.13.1 for details).
There are many more role particles than just the basic ones listed above. Most of the items are ‘complex particles’ formed by concatenating two or more words (see section 2.2 for discussion).
In elements formed by connecting two or more non-clausal elements of the same or similar type (NP, PP, ADVP, CP-THT, etc.) into coordinated structures, the connective particle P-CONN usually appears. In these contexts P-CONN projects CONJP. We refer to this type of coordination as ‘non-clausal coordination’, although instances of CP are clauses.
鈴木さん と 高津さん が 協議 し た 。
スミス か ジョーンズ か アンダーソン が 契約書 に サイン し た 。
机 や 椅子 を 並べ て 下さい 。
捕り方衆 の 叫び声 が あっち から も こっち から も 聞こえ て 来る 。
しかし 運慶 の 方 で は 不思議 と も 奇体 と も とんと 感じ 得 ない 様子 で 一生懸命 に 彫っ て いる 。
In addition, various conjunctional adverbs (CONJ) can appear in the same environments:
近代以降 は 女子 あるいは 成人女性 が 舞う 場合 も 多い 。
外国語 は 、 楽しく 、 かつ 効果的 に 勉強 し たい 。
See the discussion on the pattern of non-clausal coordination in section 3.5.
Linking phrases of the same (or similar) type of clauses into a series can create different sorts of structures. For IP-ADV instances, the linkage can be either subordinating or coordinating, and relative clauses modifying the same single head N are linked in parallel, simple parataxis (for this latter case see section 3.2.6.1).
Among clause-linking connective particles (P-CONN), there are those that take an adverbial clause (IP-ADV) as their complement (projecting a PP), and there are those that attach directly to the predicate of the adverbial clause (placed under IP-ADV). In the first scenario, if the clause is subordinated, -SCON will be added as an extension forming PP-SCON (possibly with additional extensions, e.g., PP-SCON-CND for a subordinate clause that is a conditional), whereas a coordinated clause will appear under PP-CONJ. In the second scenario, depending on the type of clause linkage, one of these three extensions will be added to the clause, forming either IP-ADV-SCON (also IP-ADV-SCON-CND), or IP-ADV-CONJ. The specifc connective particles are introduced below. For details on control environments in adverbial clauses, see section 5.3.1. For details on the distinction between subordination and coordination, see section 5.1.5.3.
In scenarios where connective particles (P-CONN) link non-final clauses (IP-ADV) to following clauses, if the linkage is subordinating, the function of the subordinated clause is sometimes conditional, and the constituent will take the extension -SCON-CND.
試験 が あまり 難しい と 、 合格者 は 出 ない だろう 。
Note that it is not always the case that (P-CONN と) introduces a conditional clause. For example, in (2.72) below a non-conditional subordinate clause is marked by this element:
太郎 が ギター を 弾く と 春子 が 歌っ た 。
Furthermore, (P-CONN と) can mark coordinated clauses as in the example below:
英語 が 上手 だろう と 下手 だろう と そういう こと は 関係 ない 。
Note incidentally that (P-CONN と) can coordinate non-clausal constitutents, and that (P-ROLE と) can express grammatical roles, and that (P-COMP と) marks quotations.
When a connective particle P-CONN is analyzable as a having the character of a suffix, it is placed after a stem as the last element of an adverbial clause (IP-ADV), and the extension specifying its function (e.g., -CND) is attached to the IP-ADV node. Below are examples of conditional connective particles of this type.
だって 携帯 の 会社 変え たら アドレス とか 全部 変わる もん ね 。
まっすぐ に 王城 に 行き着け ば 、 それ で よい の だ 。
ビール は よく 冷え て い て も 飲み たく ない 。
Note that for the forms listed as connective particles (P-CONN) above, in addition to their use in PP-SCON-CND and IP-ADV-SCON-CND, there are also instances where the same forms appear with non-conditional subordinate clauses (PP-SCON and IP-ADV-SCON), and some of these are homophonous with non-connective particles as well. For example, the connective particle たら in (2.77) below introduces a subordinate clause, but it is not the case that たら marks a conditional clause here.
奇体 だ と 思っ て い まし たら 、 また 腹かけ から 何 か 出し まし た 。
There are also cases where connective particle ば doesn't mark conditional clauses. Below is an example of ば marking a coordinated clause ば.
漆 も 塗っ て なけれ ば 磨き も かけ て ない 。
For linkages between clauses, the other type of subordinating conjunction is non-conditional. The semantic effects of the subordinated clauses are various: concessive, adversative, purposive, cause-introducing, grounds-introducing, reason-introducing, additive, alternative, expressive of temporal relations of concurrence, simultaneity, priority, etc. In the configuration involving particles that aren't analyzed as suffixes, a dependent clause IP-ADV is the complement of a particle phrase PP with the extension -SCON added to that PP node.
この 論文 は 2 度 読ん で み た けれども 、 理解 でき なかっ た 。
用事 が あり ます から 、 失礼 し ます 。
毎朝 ご飯 を 食べ て から , コーヒー を 飲み ます 。
この 薬 は にがく ない ので , 飲み やすい です 。
Connective particles ながら, たら/だら, and て/で can mark adverbial clauses expressing temporal relations (simultaneity or priority) as well as cause, grounds, and reason. These particles are placed under an IP-ADV, to which the extension -SCON is added. Below is an example with ながら expressing simultaneity:
太郎 が ギター を 弾き ながら 歌っ た 。
For linkages between clauses, another relation mediated by particles is that of coordination between independent IP instances. In coordinated clause linkage with a non-suffix connective particle, the preceding clause is an IP-ADV under PP-CONJ. In coordinated clause linkage with only a suffix connective particle under the IP-ADV, the extension CONJ attaches directly to the IP-ADV node. When more than two clauses are coordinated, the structure is different from that for subordinated clauses in that each non-final coordinated clause appears directly below the clause that immediately follows it. For the purposes of semantic interpretation, however, clauses so coordinated are treated as being at the same structural level. See section 5.4 for a discussion of the conditions on antecedents for co-reference with empty positions in this configuration.
Here is a list of some of the particles that mark coordinate IP instances:
There is some overlap in the inventory of coordinating and subordinating particles. Only a few particles are regularly coordinating in clause final position. Whether a clause is independent of and not embedded within an immediately following clause or not is decidable by the meaning of the sentence as a whole and by the patterns of constituent-sharing possible between clauses. (see section 5.1.5.3 for details). The example below has a connective particle が following the first clause. Note that the subject argument position for both clauses takes the same antecedent (here a zero pronoun in a leftmost ATB position).
前回 は 民主党 の 推薦 を 受け 大勝 し た が 、 今回 は 市民党 の 立場 で 戦う 。
Connective particles て/で, たり/だり, たら/だら, ば, and たって introducing non-final coordinated clauses appear inside those clauses rather than following them. In such cases the extension CONJ is added to the IP-ADV node. In the example below, a null-pronoun subject and three overt constituents are shared by the coordinated clauses.
公園 で は 肩掛け送風機 で 落ち葉 を 吹き寄せ て 一 箇所 に 集め て いる .
Complementizer particles such as と or という (P-COMP) introduce a clausal constituent that denotes the content of speech, thought, representation, or reception which may be anything from a sequence of full sentential utterances to an interjection or a fragment. They take as their complement various phrases such as IP-SUB, CP-QUE, IP-IMP, CP-FINAL, CP-EXL, INTJP, multi-sentence, IP-SMC, etc. and project the phrasal category CP-THT, which in turn serves as the complement of a predicate or a noun (See section 6.18 for details). In addition, the particole と takes NP as its complement and projects PP-CMPL.
Below is a list of some of the complementizer particles introducing content complements of predicates:
ジョン は 、 ビル が 自分 を 傷つけ た と 言っ た 。
Below is a list of some complementizer particles introducing content complements of nouns:
ギリシャ語 の 忘我 ( エクスタシー ) は 何 か の 横 に 立つ という 意味 です
Some complex particles like という and との allow questions to modify nouns:
こんな こと を 続け て も いい の か という 疑問 が 絶えず 私 を 苦しめ た 。
Note that when introducing complements of nouns, if particles such as という and って take an NP rather than a clause, then those particles project a PP not a CP:
晩年 の 政宗 は 、 『 酔余口号 』 という 漢詩 を 残し て いる 。
In naming constructions with verbs such as 名付ける or 呼ぶ, the extention -CMPL is added to PP projected by P-COMP と (that appears with proper nouns or other ‘mention’ uses of words).
オンラインコミュニティー の 中 に 、 「 伊達fan 」 と 名付け た 特設ページ を 設ける 。
あらゆる 自然災害 に 備え 、 暮らし の 安全安心 を 生み出す 事業活動 を 「 減災産業 」 と 呼ぶ 。
Utterance-final particles (P-FINAL) include the following forms. (These are more closely related to speech act type than to predication.)
These are analyzed as projecting a CP layer and as taking preceding dependent clauses (IP-SUB) as complements. The type of CP depends on the combination of final particle and IP-SUB, forming either questions (CP-QUE), exclamations (CP-EXL), or utterances not characterizable as either sub-type (CP-FINAL).
「 この 船 は 西 へ 行く ん です か 」
「 ああ 、 なんと 彼 を 愛し て おら れ た こと か 」 。
それ は 残念 です ね 。
When there are multiple sentence-final particles, they all appear side by side under the same CP projection.
「 多分 、 あなた は 食用豆類 を いく つ か 見つける こと が できる わ よ 。
Interjectional particles (P-INTJ) have a fairly free distribution, along with introducing vocative phrases. Here is a list:
The class of toritate particles (sometimes called fuku-joshi, lit. ‘adverbial particles’) are labelled P-OPTR. They can be roughly described as relating a constituent to the domain of discourse. For example, は marks things that are already in the domain, or marks elements that are in contrast to something else in the domain. And も in affirmative contexts marks things additional to a restricted domain. The distribution of toritate particles is fairly free. When a toritate particle marks an NP, ADVP, or IP, it projects a PP. When a toritate particle marks a constituent that is headed by a particle (typically a PP or a CP-THT), it appears alongside that particle.
Relevant particles include:
その 時 は 母 も 笑っ た 。
子供 は 何 と も 云わ なかっ た 。
解決策 は 当事者 しか 導き出せ ない 。
A は-marked noun phrase that has no clearly defined function within the sentence is usually labelled PP-TPC.
自分 も 確 に これ は 死ぬ な と 思っ た 。
帰り は 私 が 先生 を お送り し ます 。
However, if the role of a は-marked PP can be specified, then a dismabiguating extension such as -SBJ or -TMP is added.
今 は 国産的 な もの は 安い 。
The first subject in a so-called double subject constructions, which is usually marked by the particle は, is annotated as PP-SBJ, while the second subject (usually marked by the particle が) is annotated as PP-SBJ2.
彼 は 財布 が すっからかん だ
Sometimes particles are directly dominated by an IP without projecting a PP. Below is a discussion of the basic scenarios for this annotation.
As already pointed out in sections 2.6.1 and 2.6.2, among the connective particles (P-CONN) there are those that are best analyzed as suffixes, and these attach not to IP-ADV, but to the final element inside of IP-ADV. Below is a list of the particles of this type.
Below are examples for つつ, たって/だって, and たり/だり.
太郎 が 新聞 を 読み つつ ご飯 を 食べ て いる 。
あの 人 に は 話し たって 分かり ませ ん 。
彼女 は そわそわ と 行っ たり 来 たり し た 。
Connective particles and toritate particles can appear within complex predicates involving voice or aspect forming expressions such as [VB て/で いる/ある/おく], [VB て/で あげる/やる], [VB て/で いく/くる]. In these scenarios connective particle て/で is placed directly under an IP in a flat configuration. The auxiliary verb following connective particle て/で is labelled VB2 (see section 4.5 for details).
あいつ は きっと デート を し て いる ん だ 。
The connective particle たり/だり in expressions like [VB たり/だり する] is sister to the following VB2 する in a flat configuration. The connective particle たり/だり frequently appears in paired expressions of the form [VB1 たり/だり,VB2 たり/だりする] where the first occurrence ends a non-final clause and the second occurrence is part of a complex predicate, as in the example below (see (2.105) for another example).
うち の スタッフ が 殴ら れ たり 、 罵倒 さ れ たり する こと も あり ます 。
The same sort of annotation is made for toritate particle など in construction with a following する in expressions like VB などする. The same applies for connective particle なり and する in expressions like VB なりする.
暑い 日 は 帽子 を かぶる など し なさい
読め ない 漢字 は 辞書 を 引く なり 、 日本人 に 聞く なり し なさい 。
In negated expressions, toritate particle は frequently appears with NEG.
岩 で は ない 、
Furthermore, the toritate particles は, も, しか, ばかり, さえ, etc. appear within complex predicates immediately following the connective particle て/で and the gerund form で of the copula だ, directly under an IP.
女子 は 遊ん で ばかり い ます 。
さっき まで の 絶対的 な 自信 は やや ゆらい だ よう で は あっ た が 。
Likewise, the role particles (P-ROLE) を, に, が as well as the conjunctive particles (P-CONN) と, ば which constitute complex predicates such as [VB ざる を 得 ない], [VB etc. に 違いない], [VB わけ に (は/も) いか ない], [VB が いい/よい], [VB と いい], [VB と いけ ない] [VB なけれ ば なら ない] are placed directly under an IP.
結果 、 江戸幕府 は 院政 の 存在 を 黙認 せ ざる を え なく なる 。
その 国 は たいへん 美しい に 違いない 。
この 機会 を 逃がす わけ に 行か ない の だ 。
その 身代り を 呼ぶ が よい 。
▼ 埋蔵資源 が ある と いい 。
明日 の 朝 は 6 時 に 起き ない と いけ ない 。
生活費 は かせが なけれ ば なら ない 。
There are some instances of verbal nouns (VB) being marked with role particle を. In this scenario (P-ROLE を) is placed directly under an IP, sister to the VB that it follows.
そして 、 ある 定め られ た 日 に は 床の間 で 彼ら を 供養 を する の です 」
When a particle such as とか, と, やら, か, or も ends a non-clausal coordination, the coordination-ending particle is placed under the NP dominating the conjuncts:
この 近く に は 温泉 とか 野球場 とか が あっ て 、 とても 楽しい です 。
Aside from instances of coordination, the practice of burying P instances inside an NP is also adopted when か and も are part of an NP with an indeterminate (WH) element (see section 3.3.1.3 for details).
誰 か が 助ける だろう 。
誰 も その 問題 を 解け なかっ た 。
Instances of PP usually do not stack up. That is, a PP is normally not the complement of a P. When there is more than one particle attached to a noun, they are placed in series under the same PP.
その カート は 都市 へ と 方向転換 し た 。
この 点 は どの 作物 について も 言える こと で あり ます 。
脇役 が 良い から こそ 、 主役 が 引き立つ .
そんな に 頼ま れ て は いや と は 言え ない ね 。
The sole exception to this policy is when PP instances are coordinated (see sections 2.5 and 3.5 for details).
「 …… 卑猥 に も 不潔 に も なじむ こと が ない 。
There are instances when for a given NP, が, を, へ, に and other particles that might be expected to appear given the grammatical role played by that NP in its context do not in fact appear. This happens particularly often when argument NP instances are introduced in conversation and the headlines of newspaper articles. When such a bare NP appears at the left periphery, it frequently takes on the discourse function of a topic. When the grammatical function (subject, primary object, etc.) of a bare NP with respect to a predicate is retrievable, the annotation either specifies the function directly with an extension to the phrase tag, or by referring to a particle that would otherwise denote that function. For convenience we call these particles ‘omitted particles’. If the bare NP is an argument (i.e., has a core grammatical role), then there is no reference to a particle and the argument is directly labelled with its role, e.g., NP-SBJ, NP-OB1, etc. In (2.129) below, the noun phrase 掲示板のポスター has the core grammatical role of -OB1 with respect to the predicate 見た, and so is labelled as such.
なんか 、 美奈子 、 掲示板 の ポスター 、 見 た ?
If the bare NP is an adjunct (i.e., has a peripheral grammatical role), then a PP projection is created and a particle specifying that role is inserted as a terminal beginning and ending with ‘*’. For an instantiation of this treatment, see in (2.130) how the bare adjunct NP 炊き出し準備 is marked up as an NP complement to an omitted particle *で*:
炊き出し準備 、 高台避難後 に 自宅 に 戻る ( 宮城県七ケ浜町 )
The phrase is an intermediate level of organisation between a word and the clause. Phrases can be characterised partly by their external function and partly by their internal form. The focus of this chapter is on describing the form of phrases, although function is occasionally mentioned, typically with reference to sections of other chapters. By “form” is meant the way the structure of the phrase is made up of other, smaller elements. Each phrase is made up of a HEAD word and any DEPENDENTs. The head word is the word that the phrase is required to have in order to exist, and may be the only element of the phrase. A phrase is named after the word class of the head word. Phrases relevant for the description of Japanese are: noun phrases (section 3.2), quantifier phrases (section 3.3), and adverb phrases (section 3.4). The chapter also addresses the annotation of phrase coordination in section 3.5.
Noun phrases (NP) are canonically headed by a nominal element (e.g., N, NPR, PRO, WPRO, etc.). Q and NUMCLP can also head NP instances.
Noun phrases typically recieve role information from placement within a PP projection. But noun phrases can also be immediately dominated by IP and such noun phrases need to be marked for function. These noun phrases are discussed in section 1.6.2 which notes implicitly marked noun phrases with core grammatical roles and section 1.7.2 which discusses noun phrases with implicitly marked adjunct roles. See also 4.16 on predicate NP instances. Besides these roles, noun phrases that are outside PP may have two other roles: NP-POS and NP-VOC, which are discussed in this section.
A possessive NP (NP-POS) is an NP projected by the pronoun わが(我が). Since it combines directly with the following noun, without an intervening particle, (NP-POS (PRO わが)) is sister to the noun.
母親 は 我が 子 に 駆け寄っ た .
Vocative expressions are labelled NP-VOC.
「 先生 、 もう お忘れ です か ? 」
『 芳一 ! 』
…… 芳一 、 まア 喜べ !――
See 5.2.3 for discussion on NP-VOC in imperative sentences.
An NP without an overt head is assumed to have an empty nominal head. This situation is seen with what is sometimes referred to as N-bar deletion (or ‘truncated possessive phrase’, see 6.3), where the empty head is not explicitly included in the annotation, but is implicit in the NP that it projects.
三 番目 の に も 仁王 は い なかっ た 。
The class of common nouns (N) is the largest in the language. Here is a list of a few of the most frequently attested nouns. Note that many are ‘formal nouns’ and ‘relational nouns’.
Note also that formal nouns that follow core predicates, extending those predicates are labelled not as N, but as FN (see section 4.3 and 4.21).
Proper nouns (NPR) are the second largest class. Here is a list of a few of the most common proper nouns in the corpus:
NP layers headed by overt pronouns (PRO) (and by the under-specifed null element *pro*) are sometimes supplied with sort information. See section 1.8 for details. Here is a list of common pronouns (PRO):
Reflexive pronouns are not differentiated from other pronouns in the annotation and labelled PRO. The list of principal items is not too long, as is shown below. When a reflexive pronoun is used appositively to another noun or pronoun (as in 彼ら自身), the reflexive pronoun (PRO) projects an NP, which is placed under PRN, which in turn is sister to the noun or pronoun to form an NP (See section 1.3.8 for details on PRN). In addition, the possessive pronoun わが(我が) gets a special treatment (See section sectionref:[__possesivepronouns__]).
「 彼ら 自身 の 民族 。
Instances of WPRO are typically used in interrogative sentences. Here is a list of common WH-pronouns (WPRO):
WH-pronouns can head instances of NP suffixed by a P-OPTR か, も, or でも, to form a constituent with a quantified meaning, for example an indefinite meaning such as だれか (somebody), a universally quantified meaning such as どれも (every one), or a free-choice meaning such as いつでも (whenever). See section 3.3.1.3 for details. In general, for any indeterminate element (WPRO, WD, WADV, WNUM), a phrase that contains that element can either receive question focus or be quantified in the same way as with NP instances headed by WPRO.
誰 が 買っ た チョコレート も おいしい 。
An item Q can appear heading a noun complement, a floating quantifier, a referring expression, etc. For a discussion of quantifying expressions, see section 3.3. Here are examples of some common quantifiers (Q):
Quantifying expressions containing a numeral (NUM) or indeterminate numeral (WNUM) are labelled NUMCLP. Instances of NUMCLP are typically composed of either a NUM or WNUM paired with a classifier (CL), but there are examples having only either a NUM or WNUM. See section 3.3.1.2 for details.
Outside of the non-clausal coordination scenario (see section 3.5 for details), NP layers do not stack up in this annotation scheme. When there is an absolute need of having an intermediate layer of structure inside a noun phrase, NML is used. This only occurs when either (i) there is a modifier which modifies a coordinated element, or (ii) when a coordinated element is followed by an appositive quantificational phrase. Below is an example of the first scenario.
/ 意思表示 の 時期 ・ 方法 探る
The second scenario is exemplified below:
中立的 な 立場 の 素粒子物理学者 や 経済学者 ら 10 人 程度 で 構成 し 計画 を 検証 する 。
Simple expressions that specify the denotation of an NP by relating that NP to the discourse domain are called determiners. Determiners are labelled D, while indeterminate determiners are labelled WD. Determiners do not head phrasal categories, but rather appear directly alongside the head that they modify. Here is a list of some common determiners, some of which are attested in the corpus:
Here is a list of some common WH-determiners (WD):
A determiner D or WD is not normally a head of an NP, but appears directly under an NP as a single-word modifier of its head.
この 点 は どの 作物 について も 言える こと で あり ます 。
However, there are some cases where a D or WD is annotated as the NP head. This analysis is applied for この/その/あの/どの in このくらい/そのくらい/あのくらい/どのぐらい. As can be seen in (3.11), the NP headed by the WD どの is followed by the P-OPTR くらい.
ロンドン で は どの くらい 雪 が 降り ます か 。
In (__Kainoki_18_law_h15A119__), the NP headed by the D その is the subject of the VB 取り扱う.
___その 取り扱う 個人情報 の 量 及び 利用方法 から
Note that constituents containing WD behave similarly to those containing WPRO with respect to question focus and to quantification.
The class of prenominals (PNL) is comprised of those rentaisi that don't function as determiners. These do not normally head phrasal categories, but rather appear directly under NP. Here is a list of some common prenominals, some of which are attested in the corpus:
Although none of the items above can be used as predicates in a main or adverbial clause, some instances of PNL (notably おおきな, ちいさな, おかしな) can be modified by adverbial phrases or can themselves head relative clauses modifying a noun corresponding to the NP-SBJ in a double-subject construction: 目が大きな子犬. In these cases PNL is placed directly below IP-REL.
自分 で も 驚く ほど 大きな 声 が 出 まし た 。
For details about the phrasal category PNLP, see section 3.2.6.6.
Adnominal clauses are headed by adnominal predicates, and modify a following noun phrase head. With the exception of the short forms of the copula and some fossilized forms, adnominal precidates are indistiguishable from the past and non-past conclusive forms of predicates. Adnominal clauses are divided into three main types: gapped, gapless, and complementizer clauses.
In the remainder of this section we look at these three main types of adnominal clauses. There are also additional sections (sections 3.2.6.4–3.2.6.6) that consider further distinctions in adnominal clause type.
In an IP-REL clause a trace *T* typically occupies a constituent-initial position (see section 1.5.5 for details about trace placement) and is accompanied by information to specify a grammatical role. In (3.14), the trace has subject grammatical role (NP-SBJ *T*):
彼 は 紙 で あふれ た 机 に もたれ て いる 。
An example in which the trace has the grammatical role of primary object:
何 か 欲しい もの は ない の ? 」
An example of a long distance dependency, where the trace has the grammatical role of primary object in a clause embedded inside the relative clause.
おばあさん に 作っ て もらっ た 「 きび団子 」 を 持っ て 出発 し まし た 。
Here is an example of a relative clause in a noun phrase embedded in a second relative clause in a noun phrase embedded in a third relative clause:
これ は 、 ジャック が 建て た 家 に 置か れ た 麦芽 を 食べ た ねずみ です 。
Two relative clauses each modifying a nominal element independently of the other:
最も 貧しい 10億 人 を 表し た 向こう に ある 箱 が 見え ます か ?
An example where the modified noun corresponds to positions in each of two coordinated clauses. In this annotation, the superordinate IP-REL is given a trace, and this is inherited by the subordinate clause through ATB extraction.
彼女 は ピアノ が 弾け そして 絵 も かける 人 です 。
In the second group of adnominal clauses, the modified noun has no direct grammatical role with respect to the predicate of the directly modifying clause, which itself is a saturated structure. This type of ‘gapless’ adnominal clause is assigned the label IP-EMB.
One common type of IP-EMB construction involves the modification of unsaturated nouns. Given that many unsaturated nouns (こと, ため, の, よう, etc.) rarely appear in unmodified form, and then only with particular meanings (為になる話; 事が運ばない; モノを言うのは金だ; etc.) it is clear that in a usage as an unsaturated noun, an element cannot ‘reconstruct’ into its modifying clause and is modified by IP-EMB.
かけぶとん を はねのける の は 、 まったく 簡単 だっ た 。
沿線 に は 大学 が 多い ため 学生利用 が 多い 。
私 が する よう に やっ て み て ください 。
In general, nouns that can bear propositional content all can take saturated adnominal clauses: Such clauses modifying nouns denoting forms of speech, writing, thought, perception, sense, event, activity, disposition, cause, outcome, manner, depiction, etc. are labelled IP-EMB. The class of ‘picture nouns’ is included here.
子供 が 笑っ て いる 写真 が 置い て あっ た 。
Additionally, when a ‘relative’ noun (Teramura-E 1984, Nihongo 2009) (e.g., 時, 間, 前, 後, etc.) is being used as an unsaturated noun and takes a saturated adnominal clause, that clause is also tagged as an IP-EMB.
画面 が 見え なく なる 前 に
Clauses preceding and modifying 以来 and 以降 are headed by verbal predicates in the テ-form, but these are labelled IP-EMB as well.
肝臓 を 患っ て 以来 、 酒 は やめ て いる 。
It is possible to consider a noun like 時, 際 or 場合 appearing with a modifying clause as having some sort of adjunct-like role within that clause, warranting the use of IP-REL, but the annotation policy here emphasizes the continuity these words have with relative nouns such as like 前 or 後, and adopts the IP-EMB analysis for the modifying clause, in principle.
今度 彼 に 会っ た とき に これ を 渡し て 下さい 。
Of course, when a grammatical role is clearly involved, an IP-REL is adopted:
そして 意外 な 時 に 出 て 来 て 外界 を のぞく 事 が ある 。
Conversely, when there is no clear grammatical relation between a modified noun and the clause that modifies it, the clause is treated as an IP-EMB.
魚 が 焼ける におい が する 。
A CP-THT headed by a complementizer (P-COMP) such as という or との can also modify a noun.
外国人 に 日本語 を 教える という 仕事 は 容易 な 仕事 で は ない 。
この 計画 の 狙い は 何 か との 質問 が あっ た 。
Internally headed relative clause constructions are extremely rare, but when they are found, the annotation pratice is to add (PP (NP *T*) (P-ROLE *.e*)) immediately to the left of the constituent (either NP or PP) that is co-referent with the modified head の:
財布 が 落ち て いる の を 拾っ た 。
それから 星 の 破片 の 落ち た の を 拾っ て 来 て 、 かろく 土 の 上 へ 乗せ た 。
When a modified N has a peripheral grammatical role with respect to the predicate heading an immediately modifying IP or is moved from an adnominal constituent (e.g., a possessive PP, a relative clause etc.) within an NP with a grammatical role, a pronoun or referring expression co-referential with the head can sometimes appear. This is a special case where the modifying IP is marked as IP-REL and the trace is put into a bare PP headed by (P-ROLE ***). Situations such as these are annotated with identical sort information on the co-referential items:
有間皇子 が どう し て も そこ から 逃れる こと の でき なかっ た 悲運
自分 の 兄弟 だけ が その 中 で 苦しん で いる 壕 の なか
The category of PNLP is reserved for noun modification which cannot be adequately described according to any of the following scenarios:
For example, there are uses of one-place イ-adjectives in noun-modifying contexts where the target of modification N is not the subject of the イ-adjective in any direct sense, and the modifying element does not denote the propositional content of N.
昨晩 古い 友人 が 訪ね て 来 た
Note that PNLP instances are not headed by a PNL. PNL instances are not themselves modified, and do not appear in predicate position in main clauses or adverbial clauses, but, except for some cases, modify a given N by directly attaching to the NP which N projects (See section 3.2.5).
This section looks at the family of expressions that are headed by elements used for quantification. These elements are broadly divided into three categories: (i) elements that are quantifiers in their own right and do not contain numerals, (ii) elements that involve numerals, and (iii) indeterminate (wh-) words that function quantificationally when paired with the particles も and か. These preliminary elements are henceforth named ‘quantifier radicals’. Quantifier radicals immediately project NP layers, sometimes with modifiers included inside. These NP projections will be called ‘quantificational NPs’. Quantificational NPs can be either independent or further take host nominals as their quantificational restrictors.
Words that form quantifier radicals without numerals (NUM or WNUM) are labelled Q (quantifier). Q instances always project an NP layer.
泥棒 は 店 から 宝石 を すべて 奪い去っ た .
Below is a list of frequent Q items in the corpus:
Quantifying expressions which contain a numeral (NUM) or indeterminate numeral (WNUM), are labelled NUMCLP regardless of whether it is ordinal or cardinal.
青春時代 、 友達 が デート __に 明け暮れる の をよそに 、 僕 は 家 に 一人 こもっ て SF を 読ん で い た 。
この アート は 何百 年 も 続い て いる の です
第一 章 総則
NUMCLP instances are typically composed of a NUM or WNUM paired with a classifier (CL). However, there are two exceptional kinds of NUMCLP which lack CL. One is such that they are solely projected by instances of (W)NUM.
私 は まだ 十八 です わ 。
もう 六 時 半 で 、 針 は 落ちつき払っ て 進ん で いく 。
The other cases involve conventional phonological idiomaticity (juku-ji-kun, 熟字訓). 一人 (hitori, one person) and 二人 (futari, two persons) are typical examples.
第2テノール の 一人 は パヴァロッティ だ 。
The immediate conclusion is that it is the NUM that obligatory appears in a NUMCLP as a head.
Classifiers are basically nominal in nature. Here is a selection of typical CL instances:
The CL word class, in fact, does not form a closed set; Japanese exhibits a flexibility that ordinary nominal expressions can also pretend to be instances of CL, as is shown by the following example (3.43) (That 特別委員会 is an ordinary noun can be verified by its possible occurrence in an argument position. Compare *個が開かれた with 特別委員会が開かれた).
次 に 、 ただいま 決定 いたし まし た 特別委員会 を 除く 四 特別委員会 につきまして は 、 理事会 で 合意 いたし まし た とおり 設置 する こと と し 、 本日 の 本会議 において 議決 する に 御異議 あり ませ ん か 。
Despite their inner complexity, NUMCLP instances behave as atomic instances of N. They further project an NP just as do quantifiers (Q). Modifications on numeral expressions are done in between the NUMCLP and the higher NP projections.
この 店 は 朝 7 時 から 営業 し て いる 。
この 3 年間 は 消費 が 冷え込ん で いる .
がらんと し た 旅館 の 一 室 。
各店 も 対策 に 追わ れる 1 年 に なる だろう 。
Expressions containing indeterminate words (includeing WPRO, WD, WADV, and WNUM) are quantifer radicals when they are accompanied by particles か and も. For convenience we refer to these as ‘w-expressions’.
If a w-expression is under a category marked by も the whole quantificational NP takes on a quantificational force of concession (free choice).
試験 に 通り さえ すれ ば 、 何 も 問題 は ない 。
会計士 は 誰 も 会議 に 出席 し なかっ た 。
津波 は 誰 に も 止め られ ない 。
If a w-expression is under a category marked by か the whole category takes on an indefinite reference with existential quantificational force.
社員 が 誰 か 移動 を 希望 し て いる 。
There are also particle combinations such as 何もかも which receive the free choice interpretation. In the NPCMJ, such complex expressions are treated as one word and labelled Q.
「 何もかも が 失敗 だ 、 ワトスン
Quantificational NPs sometimes get coordinated, the whole being instances of CONJP (see section 3.5). In the example below, the disjunctive coordinative particle (P-CONN) か connects the first numeral NP 二晩 and the second 三晩.
たとえば この あいだ も 二 晩 か 三 晩 かかっ て 小さな 額ぶち を つくり まし た 。
The example below is another coordination of numeral quantificational NPs, which is done without an overt heading P-CONN.
つづい て 和尚 は 、 第二 、 第三 、 第四 の 抽斗 を 開け た
The cases listed above make genuine coordinations from the viewpoint of semantics.
On the other hand, sequences of quantifier radiclas that can't be analyzed as genuine coordinations are juxtaposed as they are. They will share a single NP projection together. Dates and time make typical cases of this.
さわやか な 九 月 一日 の 朝 でし た 。
六 時 半 に W君 が 来 た 。
Below is another case of non-genuine coordinations in which a numeral NP consists of a duplicated NUMCLP, duplication of which emphasizes distributivity and totality.
「 それ を 一 枚 一 枚 はがし て 、 貧しい 人 に あげ なさい 。
A quantificational relation between a quantificational NP and another NP is said to be established if and only if the quantificational NP specifies either the quantitity or the proportion of the other NP (i.e., the other NP makes a quantificational domain). The other NP is called the ‘host NP’. For example, in (3.58), the host NP of the quantificational NP 多く is 人. 多く specifies the quantity of people that 人 denotes as large.
多く の 人 が その 本 を 褒めちぎっ て いる .
Not all quantificational NPs have quantificational relations with their adjacent NP instances. (__Kainoki_414_aozora_Harada_1960__) is a clear example of a non-quantificational relation: 両腕 (a pair of arms) does not specify the quantity of 父親 (the father).
かけ て いっ た 父親 の 両腕 の なか に
To make matters worse, as will be revealed in section 3.3.5.3, the types of these relations are not always tractable from the syntactic configurations. For this reason, when there is a quantificational relation between a quantificational NP and another NP, sort information is added to make the relation explicit (see section 1.8). The general rules are:
すべて の 日本人研究者 が ノーベル賞 を 欲し がっ て いる 。
ステーキ用 の 肉 が たくさん 冷凍庫 に 入っ て いる 。
集合 は 半分 が 駄目 だ
人間 の 体 は 60 % が 水 、 魚 は 75 % 、 くらげ にいたって は 96 % が 水 だ 。
There are two attested consturctions for WH-quantifier radicals:
振り返っ て 後ろ を 見 た が 、 誰 も い なかっ た .
どの 社員 も 異動 を 希望 し て い ない 。
A w-expression together followed by a particular particle makes a quantificational NP, which may optionally be linked to another external NP as its host. In (3.66), the quantificational NP だれも is hosted by the NP-SBJ 会計士たち.
会計士たち は 誰 も 会議 に 出席 し なかっ た 。
On the other hand, The sequence [w-expression + noun + particle] contains its quantificational host inside; in fact, the N in that sequence always functions as the domain of the quantification. In this regard, the N, being inside of the quantificational NP, qualifies as the host nominal. However, for the sake of easy queries for this construction, this treebank annotates it flatly, with no resort to sort information.
私たち に は 何 の 問題 も ない わ 。
どの 取引先 から も 注文 が 来 まし た 。
In a practical sense, host NP layers of quantificational NPs induced by w-expressions are all such that drift apart, as is exemplified by (3.66).
There are ample constructions involving quantificational NPs in Japanese:
These constructions are each explained in this section:
Those quantificational NPs whose domain of quantification is indicated by their adnominal modifiers (typically genitive PP) are no different from ordinary NPs modified by instances of PP or IP-REL. No sort information is needed.
ヨーロッパ人 の 全員 が ヨーロッパ の 中 を 自由 に 移動 できる 。
In the appositive construction N Q/NUMCLP, the quantificational NP is covered with a PRN, which in turn is sister to the host noun head ((NP (N host) (PRN (NP (Q/NUMCLP quantificational expression))))).
初 の 仙台公演 で は ビゼー の 華麗 な 音楽 に 彩ら れ た 名作 「 カルメン 」 全 4 幕 を 上演 し ます 。
If the quantificational NP quantifies over what the host denotes, a mininum sort information ;* is added to the quantificational NP to mark this quantificational relation: (NP (N host) (PRN (NP;* (Q/NUMCLP quantificational expression)))). In most cases, the host nominal are referentially specific, and the outer NP either is further marked by postpositions or forms an NP-PRD.
社員 全員 が 書庫 の 鍵 を 持っ て いる 。
宮城県内外 の 自治体職員 や 企業 の 防災担当者 ら 40 人 が 参加 。
Otherwise sort information is not added. In (3.73), for example, there do not exist three referents that each could be refered to as 僕たち in its own right . Rather, there is only one 僕たち, which consists of 3 members.
ナニー が 先頭 に なっ て 、 僕たち 三 人 は ベッド の 足元 に ひざまずい た 。
When a NUMCLP is followed by another bracketed NUMCLP, the latter is analysed as a parenthetical and thus projects an NP which further projects PRN.
1926 年 ( 大正 15 年 ) に 仙台市電 が 開業 し た 。
Adnominal expressions of the form [Q + の] present the need to choose one of two different analyses, depending on whether they are quantifying or referring. A good example is the expression 両方のチーム in (3.75), which is ambiguous by itself. In one possible interpretation, the NP refers to two teams (e.g., the Giants and the Tigers), in which case the quantifier radical is quantifying. In the other interpretation, the NP refers to a (single) team of two individuals (e.g., the team managed by Masayoshi Son and Sadaharu Oh), where the quantifier radical is referring (to two persons). In most cases, one needs the help of the context to make a decision. For (3.75), the predicate 攻め合う ‘attack each other’ provides enough context to establish that the Q expression 両方 quantifiers over what the host noun チーム denotes. [Q + の] forms an IP-REL;* structure:
両方 の チーム が 攻め 合っ た .
Below are additional examples of NP instances that are genuinely quantifying:
この 音楽 は 一部 の 人々 に 熱狂的 に 支持 さ れ て いる 。
空 に は 無数 の 星 が きらめい て い た 。
On the other hand, referring Q-expressions are merely PP instances (rather than an IP-REL structure, see section 4.23) with no ‘;*’-markings (e.g., (NP (PP (NP (Q 両方) の) (N チーム))). Below is an additional example:
THE FOLLOWING EXAMPLE NO LONGER WORKS! --> NOW IT SHOULD WORK!
津波 の 爪痕 を 目 に 焼き付け 、 体験談 を それぞれ の 地元 に 持ち帰る
When an NP projected by an NUMCLP is followed by の and a nominal, there are three possible analyses, depending on the relationship between the quantificational NP and the nominal following it.
The first scenario is one where the NP projected by NUMCLP functions as a genuine quantifier. In this scenario, [NUMCLP + の] forms a IP-REL;* structure:
二人 の 剣豪 は お互い に 剣 を 突き返し た .
(3.80) is an example where the quantifying NUMCLP expression denotes proportion rather than cardinality. It is annotated in the same way as above.
この 1956 年 に アメリカ で 行わ れ た 調査 で は 30 % の 人 が 人生 が 非常 に 幸せ だ と 答え て い ます
The second scenario is one where the NP projected by the NUMCLP attributes a property to N such as size, order in a series, etc. In this scenario [NUMCLP + の] forms a IP-REL or IP-EMB without any sort information.
すべて の 四 本 足 の 哺乳類 は 四 本 足 の 動物 だ 。
不意 に 第二 の アイディア が 起こっ た 。
「 ハイ 、 あれ が 二十五 歳 の 時 の お話 で ござい ます よ 」
In the third scenario, the NP projected by NUMCLP is a referring expression. It is put under a plain PP without sort information:
運命 が 二人 の 仲 を 引き裂い た .
二 月 十七 日 の 晩 で あっ た 。
Floating quantifiers are so called because they form clause-level constituents by themselves while at the same time quantifying a clause-level host NP. They are always quantifying and always specified as such by sort information with explicit designation of grammatical roles like ;*SBJ*, ;*OB1*, etc. to indicate which NP they are associated with. In general they are associated with arguments, and they themselves are never marked with case particles.
店 に は 、 人 が いっぱい いる 。
彼女 は 夫 の 欠点 を ひと つ ひと つ 言い立て た .
When the host nominal does not appear overtly in the sentence, its presence is indicated by a zero pronoun, as in the example below. Note that there is an alternative analysis to floating quantifiers hosted by covert NP instances, namely, marking those apparently independent quantifiers as argumental NPs which are case-less. This corpus prefers the floating quantifier analysis.
部長 が 、 すべて うまく 行く よう に 取り計らっ て くれ た .
『 みな 私 の 手落ち だ !――
Numeral classifier expressions can also float.
台所 に お酒 が 二 升 あっ た 。
彼女 は ブーツ を 一 足 買っ た 。
An example of floating NUMCLP quantificational NPs with its host being covert:
すると また 一 匹 あらわれ た 。
W-expressions can also float.
「 ここ に は 、 生き て いる もの は だれ も い ない よ 、
An example of floating w-expressions with its host being covert:
巴里 の 有名 な 貿易商 、 山田和市氏 の 夫人 と 令嬢 で 、 どちら も 相当 に 日本語 を 話す 。
Floating quantifiers, numerals and w-expressions can, though not so frequnetly, precede their hosts:
しかし その 海岸 一帯 に は 、 たくさん 不思議 な 事 が 見聞き さ れる 。
「 もう 一 つ ルビー を 持っ て いき ましょ う か 」
彼ら 三 人 しか 客 が 乗っ て い ない 電車 に は 、 暖かい 陽 が ふり注い で い た 。
誰 も 消火 に 手伝う 者 は い なかっ た 。
When a Q or NUMCLP is paired with a specific host and at the same time particle-marked for a core grammatical role, unlike bare floating quantifiers discussed in section 3.3.5.5, its relation with the host is a partitive one. In this case, the Q or NUMCLP is treated in the same way as any other sort of argument, i.e., as a referring expression. The host NP is placed in a sister position to (P-ROLE ***) and marked with sort information. A null pronoun with the same sort information is then introduced in the complement position to the Q or NUMCLP: (NP (PP (NP;{WHOLE} *pro*) (P *の*)) (Q ...)).
自分たち は ほとんど が 次男坊 ばかり です から 」
県内 で 48 店舗 を 持つ 同生協 は 、 被害 が 大きかっ た 2 店 が 閉店 し た 。
製造業 は 14 業種 の うち 11 業種 、 非製造業 は 12 業種 の うち 9 業種 が 上向い た 。
Context by itself can be sufficient grounds for treating a Q expression as independently referring rather than as quantifying, as has been briefly discussed in 3.3.5.3. However, some Q expressions are interpreted as referring by virtue of elements in the expressions themselves. In some expressions, for example, a Q or NUMCLP heads a clausal constituent and takes a D modifier or a definite NP complement that makes the reference definite (see section 3.3.1.2). Such expressions can be treated as referring, and they are not linked to any hosts with sort information, as is already mentioned in section 3.3.5.1. For instance, in (3.102) below, the quantifier すべて takes a の-marked NP complement that serves to make the whole Q expression definite. There is a strong tendency for these referring expressions to be marked with particles to be sentential arguments:
国内企業 の すべて が 週休二日制 を 導入 し て いる 。
アンケート は 県内 の 沿岸自治体 15 市町 の 245 校 を 対象 に 実施 し た 。
本年度 は 2 月 1日 から ___3 月 ___10日 まで 計 約1800 ___人 が 参加 し た 。
Quantificational NP instances originated from Q (or, more frequently, NUMCLP) sometimes function as adverbs that don't quantify any host nominals. Among the typical cases are adverbs which express time, duration, distance, or degree. No sort information is given to these adverbs. Rather, these adverbs are identified as either an NP-TMP, NP-ADV, or NP-MSR (see section 1.7.2 for the discussion on the extensitons -TMP, -ADV, and -MSR):
尿道 を つくる 工程 と 似 て い ます が もう 少し 複雑 です
全国平均 を 5.3 ポイント 下回っ た 。
私 は 、 もう 少し ここ で 待と う 、 と 考え て い た 。
スミス は 2 年間 バーミンガム に 住ん だ 。
Instances of ADVP usually function to modify predicates. They can also directly modify nouns that denote amount, degree, extent, etc.
けれども 耕助 の いかり は なかなか 解け ませ ん でし た 。
学校 の 最も 近く に 住ん で いる の は 佐藤さん です 。
Adverb phrases (ADVP) usually consist of an adverb head, possibly accompanied by modifiers and/or complements. Here are a few of the more common adverbs (ADV) in the corpus. There are also many lexicalised instances of ADV that have been derived from [VB + (P-CONN て)] sequences.
Here are examples for WH-adverbs (WADV):
Instances of ADV are sometimes used with a P-ROLE. Representatives of such P-ROLE instances are と, に and の (there are also examples of として, という, で, etc.). They are sister to ADVP instances projected by an ADV and themselves project a PP layer.
すぐ に 決着 する わ 。 」
彼 は あまり の 悲しみ に 泣き叫ん だ .
There are adverbial expressions which end with と or に. They are treated as one word which is labelled ADV as long as they cannot stand on their own as words without と or に.
On the other hand, there are adverbials ending with と which drops optionally. In such cases, と is analyzed as an independent word and labelled AX, although the maximal projection remains to be ADVP.
ドール は ゆっくり と 携帯武器 を はずし た 。
息切れ が 早く も はっきり と 表われ 始め た 。
イ-adjectives (ADJI) in the infinitive inflection can head an ADVP.
僕 の 友だち は 優しく 、 寛大 に 微笑ん だ 。
The copula (AX に) or (AX と) preceded by an ADJN can also head an ADVP.
ドール は 不機嫌 に 言っ た 。
彼 は 茫然 と 、 立ちすくん だ 。
Instances of ADVP can be followed by toritate particles.
たまに は おごれ よ 。
Instances of ADVP appear sometimes in conjunctive structures. In many cases, conjoined adverb phrases are connected by means of conjunctional adverbs such as そして, かつ, and あるいは.
彼ら は 狙い を うまく 、 そして 完全 に 定め た 。
This section considers coordination between constituents that are not adverbial clauses. Such genuine coordination can occur between two or more instances of ADVP, PP, NP, CP-QUE, etc.
When phrases are connected with conjunction, all but the last conjuncts are labelled as CONJP instances, headed by CONJ or P-CONN, and adjoined to the last conjunct at the phrasal level, as illustrated in the schema of (3.120).
There are many particles that appear in patterns to form non-clausal coordinations. See section 2.5 for more discussion.
There are a number of set constructions with a coordinating function, e.g.,
A conjunctive particle (P-CONN) or coordinating conjunction (CONJ) can head a CONJP dominating a non-final conjunct NP.
昔々 、 ある 所 に おじいさん と おばあさん が おり まし た 。
まず 、 租税 及び 印紙収入 について 申し上げ ます 。
A P-CONN and CONJ can cooccur under a CONJP.
この 申込書 は 本人 か 、 もしくは 保証人 が 記入 し なけれ ば なら ない 。
A conjunctive particle is frequently paired with the same particle that marks the entire coordination, forming a set construction of the type mentioned above. The particle that marks the entire coordination indicates the nature of the coordination as either an exhaustive list or an inexhaustive list. These list-ending particles are ‘buried’ under the superordinate NP (see section 2.12 for more discussion).
こうして 天 と 地 と 、 その 万象 と が 完成 し た 。
私 は 電車 か バス か で 行き ます 。
PP instances can also form the non-clausal coordination structure. Note that particles marking the above-mentioned set constructions do not project PP layers on their own, as in (3.124) and (3.127).
終日 近鉄線内 のみ 、 もしくは 地下鉄烏丸線内 のみ の 運用 も ある 。
捕り方衆 の 叫び声 が あっち から も こっち から も 聞こえ て 来る 。
Coordinated ADVP instances usually involve an overt coordinating conjunction such as また, かつ, しかも, しかし, etc.
涼しさ の 生じる ため に は 、 どう も 時間的 に また 空間的 に 温度 の 短週期的変化 の ある こと が 必要条件 で ある らしい 。
そんな 服 を 着 た まま 、 この 老人 は ひどく 窮屈 に 、 しかし 安らか に 眠っ て いる の だっ た 。
CP-QUE instances are sometimes coordinated, either with or without overt conjunctions. When each conjunct is followed by the same particle such as か, there are two possible ways of annotation: One is to regard か as P-FINAL which projects the CP-QUE; The other is to treat it as P-CONN which forms the structure mentioned in section 3.5.1. The choice depends on the nature of the coordination. When CP-QUE instances are merely conjoined, the former strategy is adopted, as in (3.130). When they are alternatives to each other, the latter is applied, as in (3.131).
それ は 、 わたし が どこ から き た の か 、 また 、 どこ へ 行く の か を 知っ て いる から で ある 。
傘 を 駅 に 置き忘れ た の か 事務所 に 忘れ た の か 、 はっきり 覚え て い ませ ん 。
Note that in cases where multiple utterances of interrogative sentences are quoted, the tag multi-sentence is used (see section 6.6 for details).
え ? どの 医者 が そんな 便利 な 診断書 を くれる ん です か 、 と 伊沢 が 仰天 し て 訊ねる と 、 仕立屋 の 方 が 呆気にとられ た 面持 で 、 なん です か 、 よそ じゃ 、 そう じゃ ない ん です か 、 と 訊い た 。
The predicate is the most important syntactic unit in a clause, with the functions of almost all other constituents being oriented thereto. Predicates in Japanese can be extended by elements that specify voice, aspect, polarity, modality, evidentiality, etc. The basic elements of the predicate are discussed in this chapter.
When appearing with subjects, predicates formed on VB project clauses. In (4.1), the verbal syntagm 覚った heads the matrix clause (IP-MAT), and the extended verbal syntagm 達したのだ heads the complement clause (IP-SUB).
芳一 は 大きな 門口 に 達し た の だ と 覚っ た ――
家事 一切 に 関わら ず 、 のんびり と 心 の ぜいたく を する 。
冷蔵庫 に 入れ て も 硬く なら ない 不思議 な 「 バター餅 」 。
邦彦 は いったん やめ た たばこ を また 吸い出し た 。
ただ 餅 を 搗く 音 だけ する 。
Expressions that are treated as lexical compound verbs with the LUW (Long Unit Word) analysis of the the BCCWJ and CSJ corpora (e.g., かきわける, そぎとる, とりいれる, はりつける, etc.) are kept as one segment in the NPCMJ, but auxiliary verbs of various types (which are chunked in LUWs in BCCWJ and CSJ) are separated into their constituent segments in the NPCMJ and mostly labelled VB2. A traditional definition of ‘auxiliary verb’ is a form that can appear as a core lexical verb in some contexts, but is also used in secondary verb position to express a value in a grammatical category such as deixis, aspect, polarity, etc.
Aside from afterthoughts (倒置法 — see section 6.21.4), the elements that can follow core predicates in a main clause, extending those predicates, are as follows:
As mentioned in section 1.3.6, these are immediately dominated by the IP, to form a relatively flat structure. As special cases, we estabish the following categories (with accompanying node labels):
Instances of VB0 are the stems or inflected forms of verbs like する, いたす(致す), なさる, 申し上げる, 下さる, できる. Instances of VB0 appear after instances of VB that are sometimes called ‘verbal nouns’ such as 移動, チェック, etc. They too can be followed by VB2, PASS, NEG, and the like. In addition, the preceding verbal noun (VB) can be accompanied by a toritate particle (P-OPTR) such as は or も.
「 やあ 耕助君 、 失敬 し た ねえ 。 」
誰 も 殺害 に対する 責任 を 主張 し なかっ た 。
軍隊 は 一般的 な 職業 で 、 また 軍人 は 名誉 で あり 、 尊敬 も さ れる 。 」
Instances of VB followed by a VB0 sometimes accompany the prefix お or ご (that is, 御) to make honorific or humilific verb forms.
花子 は 鈴木先生 に 結果 を ご報告 し た 。
In addition, VB0 instances can sometimes appear in an honorifc form such as なさる or 下さる. Such VB0 instances must be distinguished from their VB2 (secondary verb) counterparts discussed below, using the following replacement test: First, remove the honorific prefix お/ご(御), if there is one, from the verbal noun (VB). Then replace the VB0 candidate with する. If this results in a grammatical string, then the candidate is actually a VB0 (e.g., 御安心なさい can be replaced with 安心する without losing grammaticality in (4.10)). Otherwise it is a VB2 (e.g., 置きなさい cannot be replaced with *置きする in (4.11)).
もう 大丈夫 です から 、 御安心 なさい 。
「 鉛筆 を 置き なさい ! 」
VB2s are non-initial verbs that follow lexical verbs or adjectives that normally can inflect. We present classes of this category in turn. VB2s that immediately follow a verb in its infinitive combining stem include the following:
やがて 光 は 消え 始め た 。
Secondary verbs immediately following the て-form of verbs (i.e., (VB ...)(P-CONN て) or (VB ...)(P-CONN で)) include the following items.
今 、 調べ て い ます 。
震災 は 人々 の 記憶 から 薄れ 始め て いる の か 。
Various verbs following イ-adjectives and ナ-adjectives are also analyzed as VB2:
逆 に 階段 の 段差 は 小学生 に は 高 すぎる 。
Verb ある following non-finite forms of the copula だ and following the infinitive form of イ-adjectives is also treated as VB2.
彼 は いと 高き 神 の 祭司 で ある 。
この 本 は あまり よく あり ませ ん 。
Verbs in complex expressions with deontic modal flavour are treated as VB2.
それら は 雄 と 雌 と で なけれ ば なら ない 。
「 もちろん 他 の 像 を 立て なくて は なら ない 」 と 市長 は 言い まし た 。
学校 で は 携帯 を 使っ て は いけ ない 。
Below are cases where the verb する is treated as VB2.
「 それ は 鉄 と まぜ たり 、 薬 を つくっ たり する の だ そう です 。 」
国 が 所有者 から 土地 を 一時的 に 借り上げる など し 、 復興 を 加速 さ せる べき だっ た 」
早く すれ ば する だけ 有利 だ 。
あの 男 に は 説明 し て も 分かり は し ない 。
努力 なく し て は この 事業 は でき ない 。
シカゴ に は ___親しい 友人 が 何 人 か いる ので 、 ホテル に 泊まら ず とも よい 。
なる appears as VB2 in an honorific verbal syntagm, when it appears with a honorific form of verb prefixed by お or ご (御) and followed by the particle に.
鈴木先生 は 本 を お書き に なっ た 。
When verbs する (or analogous forms いたす, なさる, etc.) and なる follow either a く-form イ-adjective, or [ナ-adjective/nominal predicate + に], frequently a small clause is projected by the predicate preceding する or なる making する or なる the first verb in a verbal syantagm, hence VB. There are three conditions that a small clause must satisfy in this context: (i) the head of the small clause must denote a property, not an event; (ii) -OB1 of する or the -SBJ of なる must be interpretable as the subject of the head of the small clause; (iii) -OB1 of する or the -SBJ of なる must precede the small clause, unless the small clause is marked with a toritate particle. If these conditions are satisfied, then a small clause analysis is adopted.
Instances of IP-SMC preceding する in the above-mentioned type of construction are given an extension tag -CNT (continuative). To instantiate, in the examples below, the realisation of the actions denoted by 細帯を長くして and 息子を言語学者にし entails that 長い is predicated of 細帯 and that 息子 is predicated of 言語学者, respectively. Accordingly, イ-adjective 長く and [言語学者 + に] project a small clause IP-SMC-CNT here.
そうして 細帯 を 長く し て 、 子供 を 縛っ て おい て 、 その 片端 を 拝殿 の 欄干 に 括りつける 。
息子 を 言語学者 に し たい なんて 。
Note that this analysis is not applied to another construction involving する preceded by an adjective: When する has no primary object (-OB1) and its subject is interpretable as the subject of the adjective, then the adjective projects an IP-SMC-OB1.
花子 は その 時 おとなしく し て い た らしい 。
Instances of IP-SMC preceding なる are given an extension tag -OB1. In the examples below, the realisation of the events denoted by 路はだんだん暗くなる and 花子はきれいになった entails that 暗い is predicated of 路, and that 花子 is predicated of きれい, respectively, so 暗く and [きれい + に] projects a small clause IP-SMC-OB1.
路 は だんだん 暗く なる 。
花子 は きれい に なっ た
Occasionally an NP-SBJ2 needs to be interpolated as a controller for the small clause in these constructions:
気 が 変 に なり そう な ん です 。
When an イ-adjective or a ナ-adjective preceding なる requires two arguments, the second argument, namely the primary object (-OB1) must be included in the IP-SMC-OB1 projected by the adjective.
おかあさん が 恋しく なっ た わけ で は ない 、
私 は また あの 花火 という やつ が 好き に なっ た 。
This small-clause analysis is contrasted with constructions such as those in the examples below, where the subject-predicate relation doesn't hold (though the matrix subject can control a position further down in the structure) and the condition on property ascription isn't met.
新しい 機種 も 気 に なる し 。
今後 の 医療運営 は 被災地 の 人口流出 に も 向き合う こと に なる 。
ADJI instances (also called イ-adjectives) such as 大きい and うつくしい have a unique conjugation paradigm which, for instance, lacks the imperative inflection. ADJI instances have some limitations on co-occurrence with modifiers and predicate extensions, and in certain contexts they exhibit special focus interactions with particles. But with regard to projection, they behave as other predicates. When appearing with subjects, outside of a noun phrase, they can project a main clause, IP-SUB under CP, IP-ADV, or IP-SMC, with basically the same constraints on grammatical function and grammatical marking as verbs and other predicates.
新市 を 軌道 に 乗せる 意欲 は 強い 。
そんな こと も ない です か ?
人間 で 言え ば 80 歳 近い のに 、 まだ やんちゃ です よ 」
家族愛 の 深さ は 尊い が 、 互い を 心配 する あまり 、 津波避難 が 遅れる 場合 も ある 。
この いちご は 甘く て おいしい です 。
阪神大震災 で も 心 の 問題 は 3 年目 に 多く なっ た 。
When an イ-adjective appears with a subject (overt or null) and modifies a noun, it projects a relative clause (IP-REL) or a content/specificational clause (IP-EMB).
ジョン は 赤い 車 を 持っ て いる 。
おそらく すでに 春 が 近い しるし だろう
There are many イ-adjectives having specific senses that require a second argument. A partial list of 2-place イ-adjectives, together with the particle(s) that typically mark(s) the non-subject argument:
Note how the イ-adjective 欲しい etc., takes two NP argument, both of which can be marked with が. The second one is assigned the role -OB1. A careful distinction must be made between this kind of sentences and the double subject sentences. If either of the が-marked NP instances, to the exclusion of the other, can combine with the predicate to build up an appropriate sentential meaning with a context, they are either the subject or primary object of the same predicate. For example, suppose that a sentence 私がチョコレートが欲しい is given, both of the sentences below are semantically well-formed and can stand alone with a context.
In this case, 私が is the subject and チョコレートが is the primary object of the sentence.
By contrast, given a sentence 象が鼻が長い (The elephant-SBJ the nose-SBJ2 is long), since in the sequence 象が長い the NP 象が cannot be interpreted as the subject of the イ-adjective 長い while preserving the meaning of the the original sentence (because it is not the elephant that is long), the whole is a double-subject sentence with 象が as -SBJ; the second subject 鼻が is assigned the role -SBJ2.
The auxiliary イ-adjective ほしい takes an IP-SMC-OB1 complement with a controllee position that regularly has a non-subject controller. The IP-SMC is a (VB ...) (P-CONN て) or (VB0 ...) (P-CONN て) sequence.
私 に 何 を し て ほしい の ?
The past tense and perfective aspect usages of た and だ are assigned to the category AXD. These follow the infinitive form of verbs. They are themselves finite inflections. The label AXD is also used for き/し, たり and り which are past tense and perfect aspect markers in Classical Japanese.
本当 に 喜べる 日 は もう 少し 先 と なっ た 。
AX is used as a label for the following elements (see section 4.13 for more details on the last case).
There are イ-adjectives that are formed with a verbal element but that end in the adjectival form づらい (e.g., しづらい, 分かりづらい, 見えづらい, etc.). These are treated as single words with the category ADJI.
Some verbal heads of IP-REL take an auxiliary た/だ, but ascribe a stative property to the modified N rather than (or aside from) participation in a past event. This only obtains in noun-modifying contexts, basically in property ascriptions. This use of た/だ is characterized by its replaceablity with ている/でいる in non-adnominal contexts without affecting the original meaning. In such scenarios, た/だ is labelled not AXD, but AX.
The following is the list of examples with た as AX.
MD, ADJI-MD and ADJN-MD are special phrase categories directly dominating words that contribute to the specification of modality for a sentence (In the following, MD is meant to stand for any of the three categories). The practice of separating out these words from others under an MD node accomplishes several things:
Most items of the category MD attach to finite inflections of predicates. Common items are:
Some examples which deserve a mention are introduced below:
彼 は 飲み物 を 勧める かもしれません 。
大昔 、 火星 に 水 が あっ た そう だ 。
先生 でも わから ない でしょう 。
その 国 は たいへん 美しい に 違いない 。
As was hinted above, there are also expressions which have something to do with modality (and which are sometimes regarded as comprising units on their own) that are nevertheless not labelled with MD. なければならない and はず are representative among them: in NPCMJ, the former retains the original conditional structure and is characterized by VB2 dominating なら; the latter is tagged as FN. See section 4.5 and section 4.21 respectively for details.
ADJI-MD is a category which dominates an イ-adjective involved in two types of modal constructions: one involves the adjective 良い and is used for expressing permission or desirability (e.g., ても良い, たら良い, ば良い, と良い) and the other is used for expressing unnecessity (e.g., までもない ). By retaining their compositional structure, the corpus captures the variability of elements participating in the constructions.
窓 を 開け て も いい ?
どう し たら いい です か 。
歴史 は 歴史家 に 任せ て 、 小説家 は 小説 を 書け ば よい の だ 。
南十字星 が 見える と いい な 。
すし が 日本 の 代表的 な 料理 で ある こと は 言う までも ない 。
ADJN-MD is a category which dominates a ナ-adjective involved in constructions expressing permission (e.g., て結構だ) or prohibition (e.g., てはだめだ ). The rationale behind their decompositional treatment is the same as for ADJI-MD.
大臣 は 退席 を さ れ て 結構 で ござい ます 。
ここ で 止め て は 駄目 だ
Most copular expressions are formed by the combination [ナ-adjective/nominal expression + copula]. However, the copula can combine with instances of PP and ADVP as well. Before going into detail about these types of expression, it is useful to consider the function of the copula, and to look at the forms and distribution of the copula. As also noted in the section 4.16, the copula functions semantically to associate a subject noun phrase with another noun in one of three basic ways:
In this section the forms and distribution of the copula are described. だ and the various simple forms analogous to だ take the role of copula and are tagged AX. The analytic form である also functions like a copula. This form is analysed into
(... (AX で)The negated copula also is analyzed, and typically a toritate particle intervenes between components:
(... (AX で)In addition to attaching to nominal expressions (NP-PRD) and ナ-adjectives (ADJN), だ and its analogous forms also attach to a clausal constituent (FN の) and some instances of MD (みたい, べき, そう).
An affirmative non-past non-volitional simple copula だ cannot appear immediately before some MD instances (e.g., らしい, みたい, かもしれない, にちがいありません, etc.) and the particle なら, whereas the complex form である can occur in this position. In the former case, a null copula (AX *) is inserted before the modal or the particle.
札 は 十 円 札 らしい 。
Volitional/presumptive copulas are analyzed as consisting of a null copula (AX *) immediately followed by a modal (だろう, でしょう, であろう).
これ は あなたがた にとって 、 小さい こと であろう か 。
An affirmative non-past non-volitional simple copula だ is omitted immediately before some sentence-final particles (か, さ). Note that in embedded indirect questions, だ can be dropped, but the sequence だか is allowed.
「 私 は 涙 という もの が どんな もの か を 知ら なかっ た 。
なぜ だ か は 分かり ませ ん 。
だ is also ommited immediately before the toritate particle なら.
「 ハンドガン なら 持っ てる よ 」
だ is optional before some sentence-final particles: ね, よ. In all these cases of だ ommission, a null copula need not be supplemented.
あなた は 女 よ 。
On the other hand, there are sentence-final particles before which だ cannot be ommitted: ぞ, ぜ, わ, な, etc. There is a modal element after which だ cannot be ommitted: hearsay そう. だ omission is also ruled out for many clause linkage particles: けれど, とも, から, し, が, etc.
Predicate extension (FN の) (AX だ) is sensitive to sentence-final particles in the same way that core copular predicates are: Long forms can appear, but the short form is regularly ommitted. Compare
Copulas have a variety of phonological shapes, as is shown in the fellowing list.
For example, the form で (in addition to supplying the stem for analytic copulas that get inflectional morphology from ある, いる, and ございます), can also appear on its own as a non-finite copula ending a non-final clause.
< SL銀河 > 客車 は 4 両 編成 で 、 定員 180 人 を 予定 。
仙台 は 静か で とても きれい でし た 。
The short form of the copula (AX で) in (4.65) and in (4.75) can in principle be substituted with であって.
Note that there is a historically attested copula なり with a combining stem なら that is identical in pronunciation to the conjunctive or toritate particle なら.
The phonological forms に and と appear in contexts where infinitive copulas are expected. Form に frequently appears after ADJN, as in (4.67), but also with nominal predicates in small clause constructions (4.68). Form と appears in the same contexts: (4.69). These に and と are analyzed as (AX に) and (AX と) in those contexts (see sections 7.2.1--7.2.3 for a detailed discussion).
子供 は 一生懸命 に 手拭 を 見 て い た 。
全学級 の 大騒ぎ に なっ た 。
急 に 室 の 中 が 暗く 陰気 と なっ た 。
Contractions of に and と with ある, なる and たる, can form adnominal copulas:
あれ は 無意味 なる 沈鬱 で ある 。
彼 に は 確固 たる 考え が ある 。
Included in the class of ナ-adjectives (ADJN, also known as 形容動詞) are so-called たる-conjugation and と-conjugation adjectives (e.g., 鬱々たる/と, 閑散たる/と, 決然たる/と, 騒然たる/と, 漠然たる/と, 呆然と, 満々たる/と, 黙々たる/と, 隆々たる/と), and so-called bungo (classical language) adjectives (速やか-なる, 確固-たる). The root form of a ナ-adjective predicate is ADJN plus its inflecting suffix AX (copula). The sequence [ADJN + AX] forms a predicate, and when paired with a subject, projects an IP. There is no doubt that these expressions can form predicates, and there is plenty syntactic evidence to show that they can project clauses. They are assigned a separate part-of-speech due to (i) their ability to take a specialized form of the copula, な; (ii) their resistance to noun-modifying expressions; (iii) their resistance to direct marking by particles other than か and complementizers と, という, といった, etc.
Predicates formed on ADJN instances project clauses in the same way as verbs and イ-adjectives:
父 は 娘 が 心配 だ
太郎 と 花子 で は 、 どちら が 歌 が 上手 です か 。
家庭 が 不安定 だ と 子ども の 状態 は 良く なら ない 。
仙台 は 静か で とても きれい でし た 。
また も 不思議 に 思っ て 彼 を 見つめ た 。
『 可哀そう な 男 だ 。
大衆 が 静か な こと は 日本人 の 特徴 で ある 。
Note that, as with イ-adjectives, when used adverbially, an ADJN followed by a copula in the ren'yookei form (に or と) projects an ADVP.
カメラ付き携帯電話 の 普及 など で 、 一般 の 人々 が 写真 や 映像 を 撮影 し 提供 する 機会 が 飛躍的 に 増え た 。
彼 は 公然 と 語っ て いる のに 、 人々 は これ に対して 何 も 言わ ない 。
Some predicates formed on ADJN instances such as 好きだ, 嫌いだ, and 上手だ take two が-marked NP instances, one of which is analyzed as having a subject role (-SBJ). The second one is assigned the role -OB1.
鈴木さん が 一番 英語 が 上手 です 。
A careful distinction must be made between this kind of sentence and the so called double subject sentence. In the first kind, given sufficient context, after removing one of the NP instances marked with が or は, the remaining NP together with the predicate form an independently meaningful sentence. In short, when the NP instances marked with が or は are either the subject or the -OB1 argument of the predicate, the construction is not a double subect construction. In the example below 必要だ is a two-place predicate with 専用アプリのインストール as -OB1, while 簡単 heads a double subject construction with スマホやタブレット端末 and 操作 as its subjects.
スマホ や タブレット端末 は 専用アプリ の インストール が 必要 だ が 、 操作 は 簡単 。
There are words similar in form to ナ-adjectives, but are traditionally put under the category of rentaishi because they only ever occur in an adnominal position, although some of them can project clauses: 大きな, 小さな, 可笑しな. In isolation, they are tagged as PNL. See section 3.2.5 for details.
In addition to the typical ナ-adjectives defined above, there is also a set of words (e.g., 不思議な/の, 当たり前な/の, キレイ好きな/の, etc.) that, while having some of the basic characteristics set out above, can also take the short-form adnominal copula の as an inflecting suffix. The copula in expression pairs such as 別な食べ物/別の食べ物 and 特別な措置/特別の措置 can be seen to alternate in form between な and の without a change in meaning. These are also put in the class of ナ-adjectives unless there is a compelling reason to do otherwise.
However, there are instances where such items have a function other than that of property ascription. Comparing 元気な子供 with 元気の源, we assign ADJN to the former and N to the latter. Furthermore, when an element that shows な/の alternation is subject to the sort of modification applied to nouns (seen in both 鮮やかなピンクの花 and 鮮やかなピンクな花) the practice is to treat that element as a N.
Note that words like よう, はず, and ふう, while accepting noun modifiers and various role particles fairly freely, take the special な-form of the copula rather than の when heading a noun-modifying expression. These are analyzed as formal nouns (FN). See section 4.21 for details.
Nominal predicates are headed by N, NPR, Q, PRO, WPRO, etc. and are assigned the label NP-PRD. A nominal predicate combines with copula to either (i) denote a property of a subject NP, or (ii) predicate an identity relation with a subject NP, or (iii) specify a member of a group-denoting NP, or value of a property-denoting NP.
As with other predicates, when appearing with subjects, predicates formed on NP-PRD project clauses:
暗黒エネルギー の 正体 を 探る こと が 現代宇宙論 の 最大 の テーマ です 。
- 審査委員長一押し の 提案 は 何 です か 。
自分 は 虜 だ から 、 腰 を かける 訳 に 行か ない 。
藤原氏 は 小保内氏 の 側近 で 、 昨年 3 月 に 退任 。
その 間 が 、 たっぷり 一 時間 は あっ た 様 に 思わ れ ます 。
国際学術コンペ の 審査委員長 で 、 フロリダ大西洋大教授 の フランク・シュニッドマンさん に 聞い た 。
だとすると 、 彼 が 犯人 で ある 可能性 は 低い 。
「 先生 、 この 人 が 生れつき 盲人 な の は 、 だれ が 罪 を 犯し た ため です か 。
Instances of NP and ADVP marked by role particles (P-ROLE) and/or toritate particles (P-OPTR) can form copular expressions, typically in pseudoclefts. In this scenario, the PP projected by the particle gets the extension -PRD to form PP-PRD.
第二 は 、 本法改正 の 遡及適用 について で あり ます 。
私 が 持っ て いる の は 本 だけ です 。
Instances of ADVP headed by adverbs can be a predicate followed by a copula. In this scenario, the ADVP projected by the adverb gets the extension -PRD to form ADVP-PRD. Typically, demonstrative adverbs such as そう are involved in this kind of construction.
ルール は こう だ 。
In a few cases, instances of IP-ADV can appear as complements of copulas. Such IP instances, too, get the extension -PRD to form IP-ADV-PRD.
視覚的 だっ たり 聴覚 から で あっ たり 、 感触 から で あっ たり 抽象的 な 捉え方 も し ます し 、 動き ながら だっ たり
Instances of CP-THT and CP-QUE can be a predciate followed by a copula. See section 6.18.6 and section 6.17.4 respectively.
Expressions headed by a [formal noun + copula] have the following grammatical functions: At the ends of main clauses they indicate various orientations of aspect, information structure, modality, evidentiality, etc. The basic members of this type are はず (plausibility); わけ (grounds, conclusion); もの (habituality, inevitability); To these can be added others that appear in a variety of contexts: ふう (mode, manner); つもり, 気, 魂胆 (intention); ところ (pre-onset, perfective, hypothetical); ため (purpose, cause); せい, おかげ (cause); ほう (alternative); まま (current state) etc. Most of these do not allow が/の alternation in subject marking of a preceding adnominal clause (unless the copula following them is itself adnominal). They are labelled FN in the corpus. Ending main clauses, they are placed directly under the IP.
こう 感じ て いる 酔客 は 多い はず だ 。
Note that for some formal nouns (e.g., はず, わけ), when they extend a nominal predicate, the copula of the predicate is the な-variant of the adnominal copula.
この 計画 は 、 実現不可能 な わけ で も ない 。
The element の that forms the のだ construction is also labelled as a formal noun (FN), and occupies the position directly dominated by IP, at the same level as the head and other constituents of IP. The primary motivations for annotating with this flat structure are (i) to avoid the difficulties in specifying scope effects that an embedding would entail, and (ii) the lack of が/の alternation in のだ constructions that would be present in an embedding.
田中 が 会い に 来 た の で は ない 。
Note that present tense nominal predicates heading のだ constructions are formed on the な-variant of the adnominal copula.
この 猫 は 、 いわば 、 我が家 の 一員 な ___の __です ___。
The distinction between the copular use of the form の and its use as a particle is difficult to make. The annotation practice here is, in principle, to view instances where [NP1 の N2] can be restated as [[NP N2] は NP1 である] or [NP1 である N2] as unproblematically copular, including instances where の indicates an identity relation between NP1 and N2. Here the form is labelled AX, with [NP1 の ] heading either an IP-REL or IP-EMB modifying N2.
「 そこ に 小さな マッチ売り の 少女 が いる 。
However, the principle above leaves room for some uncertainty. Below is an explanation of policies to resolve instances where uncertainty may arise.
被災地 で は まだ 、 保護者 が 大変 な 時期 の ため 、 子ども が うまく 甘え られ ない 面 も ある 。
彼女 は 彼 の ため に 口添え し た 。
この 「 毒殺未遂事件 」 の 正体 は 、 反政宗派一掃 の ため の 自作自演 説 も ある 。
彼女 は ガン の せい で 死ん だ 。
あなた の せい で 、 こんな こと に なっ た ん です 。
旦那さん が 本当 の こと を 言わ ない ん です か 。
どうやら 全員 ご賛成 の よう です ので 、 その よう に 決め させ て いただき ます 。
ブラウンさん は 日本人 の よう な 考え方 を する 。
ジャック は どう し て も 私 の 手紙 に 返事 を 書こ う と し ない 。
近代的 な 官僚制 の 誕生 と 産業革命 の 開始 とともに 始まっ た の です
洪水 で 3 人 の 人 が 行方不明 だ 。
However, if the complex [NP1 の] neither quantifies nor ascribes a property to N2, then NP1 is treated as any other noun phrase, and の is analyzed as P-ROLE.
彼 は 二人 の 結婚 に 立ち会っ た .
Substitution with である as a test for the status of の as an adnominal copula does not identify all the instances where a copula analysis is appropriate. Applying this test to 昔、子供の頃 yields *昔、子供である頃 which fails the test But as can be seen by 子供であった頃, the copula analysis is appropriate. An alternative test involves coordination of noun modifying clauses, where a non-final coordinated clause will end in で. Whether the target is in the non-final clause (まだ子供でウブだった頃) or the final clause (まだウブで子供の頃), by virtue of being coordinated to another copular clause, the nature of the expression in question is clear. Below is an example of this type of structure:
国際学術コンペ の 審査委員長 で 、 フロリダ大西洋大教授 の フランク・シュニッドマンさん に 聞い た 。
Clause layers (IP) are the immediate projections brought about by those predicates defined in sections 1.3.6 and sectionref:[__predicates__]. This chapter is dedicated to their subcategorization. IP is an abbreviation for Inflectional Phrase, which is another way to say ‘clause’. Instances of IP are obligatorily and unambiguously subcategorized, which is indicated by some tag extension. Table 5.1 is the full list of syntactic tags resulting from the inclusion of extension labels for subcategorizing clauses.
Table 5.1: Subcategories of IP
| Syntactic tag | Clause environment |
|---|---|
| IP-MAT | matrix clause |
| IP-SUB | clause under CP layer |
| IP-REL | relative clause |
| IP-EMB(2) | gapless adnominal clause |
| IP-ADV-CONJ | coordinating conjunctive |
| IP-ADV(2)-SCON(-CND) | subordinate conjunctive / conditional |
| IP-NMZ(2) | nominalized clause |
| IP-SMC(2) | small clause |
Instances of IP-MAT are matrix clauses which lie on the topmost layers of sentences. Basically, they are not embedded into other IP- or CP-clauses.
In principle, there is no case of IP-MAT coordinations since IP-MAT instances are meant to represent indepedent utterances, which should each be stored as a different tree record in the corpus. In the case where multiple utterances are juxtaposed under a quotation, in which the utterances (IP-MAT instances in most cases) are all placed directly under a multi-sentence phrase which is dominated by a CP-THT (sections 5.2 and 6.18). Instances of FRAG and INTJP, when forming distinct utterances, can also be placed directly under multi-sentence. See section 6.6 for discussion.
さつき町内会長 の 亀卦川正一さん ( 79 ) は 「 3 月 なら 釧路 の 最低気温 は マイナス 20 度 前後 、 昼 でも プラス に なる かどうか 。 屋外 で は 凍死 し て しまう 」 など と 話し た 。
谷藤裕明市長 は 2 月 27 日 の 定例記者会見 で 「 予定 し て い た 事業 が なかなか 進ま ない 。 何とか 工夫 を 凝らし 打開 し なけれ ば なら ない 」 と 懸念 し た 。
気迫 が 通じ た もの か 、 「 __よろしい 。 わかり まし た 」 と 相手側 の 代表者 が 意外 な 柔軟さ を みせ て いっ た 。
Quasi-matrix clauses (instances of IP-SUB) are the topmost IP clauses under matrix CP layers. They are embedded in various CP instances (see section 5.2), but never under other IP instances.
Neither IP-MAT nor IP-SUB involve control (section 5.3) or ATB-extraction (section 5.4).
Gapped adnominal clauses are tagged IP-REL and placed next to the nouns they modifiy. At least one trace *T* must be present within each IP-REL.
Multiple IP-REL instances modifying the same noun are simply placed side-by-side under the NP.
イギリス人 の 偉大 な テノール歌手 が いる 。
とはいっても 、 壁 は ぎざぎざ や とがっ た ところ が たくさん ある 念入り に 彫刻 さ れ た 家具 で さえぎら れ て い た 。
Instances of IP-REL do nothing with control (section 5.3) or ATB-extraction (section 5.4).
Adnominal clauses lacking gaps that correspond to the target of modification, are labelled IP-EMB. They typically denote the propositional content of picture nouns (走る姿), content nouns (新病棟のベッドを減らす計画), event nouns (長男を失った経験; 持ち直す可能性), relative nouns (食べたあと; 回復するはず), and function nouns (チーターの走る速度), etc.
As with IP-REL, mutiple IP-EMB instances modifying the same noun are simply placed side-by-side under the NP.
いかに も 立派 な 邸 で は ある が 、 なんとなく 様式離れ の し た 、 趣味 の 無い 、 そして 陰気 な 構造 の よう に 感ぜ られる 。
In contrast to IP-REL, an IP-EMB forms a control environment (section 5.3).
For details of IP-REL and IP-EMB as noun modifiers, refer to sections 3.2.6.1 and 3.2.6.2.
There are some exceptional instances of IP which appear in a nominal environment without any accompanying nouns or nominalizers. Such IP instances are annotated as IP-NMZ.
弟子たち は 食物 を 買い に 町 に 行っ て い た の で ある 。
しかし 幹細胞治療 を 臨床応用 する に は 長い 道のり が あり ます
Note that IP-NMZ instances are generally cases of event nominalizations in which they denote the events described by their interal predicates. This semantic property will distinguish them from targetless relative clauses (exemplified below):
残る は 勾坂甚内 だけ 。
As IP-NMZ instances are nominalized instances of IP, IP-NMZ instances are coordinated in the same way as instances of NP. Following section 3.5, NP coordinations as well as coordinations for IP-NMZ are achieved by using CONJP.
性格 が 明るい か 、 あるいは 暗い か によって 、 人生 は 非常 に 変わっ て くる 。
Instances of IP-NMZ provoke control environments (section 5.3).
In this corpus, small clauses, tagged IP-SMC, have a sense broader than the mainstream terminology, referring to clauses which are arguments of matrix predicates (which distinguishes them from instances of IP-ADV) but are with no nominal case markings (which makes them different from instances of IP-NMZ) and are deficient in terms of arguments or tenses (which differentiate them from instances of IP-SUB). Instances of IP-SMC are always selected by particular predicates specified in sections 1.6.3, 4.6, 4.8, 6.19, 6.20, 7.2.2, 7.2.3 and 7.4.
Complementizers such as と and という occasionally accompany IP-SMC instances. Such instances of IP-SMC will be immediately under CP-THT.
温度 を 一定 だ と する
Tag extensions for grammatical roles are added to IP-SMC instances when they are directly dominated by IP (4.6, 6.10, 6.12, and 1.6.3).
Coordinations of IP-SMC, if any, are treated as right node raising of matrix clauses (section 6.4). An IP-SMC provokes a control environment.
Setting aside the IP-clauses listed above, we can still find a particular kind of clauses1 that is apparently followed by another clause2, in which clause1 ends with either a non-finite inflection, a conjunctional particle, or a conjunctional adverb,
We call this kind of sentential configuration a ‘clause linkage’. In this corpus, such a clause linkage is so structured so that clause1 is dominated by another clause2, regardless of the functional linkage therebetween.
Dominated clauses (clause1) are dichotomized into subordinate and coordinate ones. In both cases, the dominated clause1 is tagged IP-ADV. Furthermore, it is placed either directly under clause2 or under a PP projected by a conjunctional particle. Instances of IP-ADV (with covering PP projections, if any) must have extension tags. The -SCON tag is for subordinate clauses, -SCON-CND is for conditional ones and -CONJ is for coordinate ones.
Instances of IP-ADV-SCON(-CND) (or PP-SCON(-CND)) are control environments (section 5.3). Instances of IP-ADV-CONJ (or PP-CONJ) allow ATB-extraction (section 5.4).
Whether multiple instances of IP-ADV-SCON(-CND) (or PP-SCON(-CND)) are placed side-by-side depends on whether or not they are each separately in a clause linkage with the dominating clause2.
Below is a non-juxtaposition example:
鶏 が 鳴い て も 女 が 来 なけれ ば 、 自分 は 逢わ ずに 殺さ れ て しまう 。
A side-by-side example:
盗ん だ 金 だけに 糸目 を つけ ず 惜し気 なく パッパッ と 使う ので どこ へ 行っ て も モテル の で あっ た 。
On the other hand, instances of IP-ADV-CONJ (or PP-CONJ), as the extension name explains itself, are dedicated to coordination with dominating clauses2. The next section will explain the way of their appropriate annotation.
For the purpose of calculating semantics and by common conventions for capturing structure, ‘clause coordination’ should mean ‘sisterhood of categories X1,2...n-1 all and only dominated by category Xn’, where X might be a full clause unit or an intermediate clause layer such as a verb phrase. But common conventions notwithstanding, the addition of an Xn layer creates a constituent for which (i) the grammatical function of material adjoining to Xn (viz. the coordinated elements X1,2...n-1) is not clear, and (ii) the function Xn has with respect to the category to which Xn adjoins (viz. the superordinate IP) is also not clear. The treebank model strives to explicitly specify a grammatical function for every constituent, so that grammatical phenomena can always be minimally defined by strctures, and thereby searched without necessitating ad-hoc navigations of unpredictable tree structure. Accordingly, a different strategy is followed that captures clause coordination as clause adjunction plus -CONJ.
In practice, coordination between a clause1 and a following clause2 is annotated so that clause1 is dominated by clause2, with the crucial addition of a -CONJ extension on clause1. Concretely, clause1 is either a bear IP-ADV-CONJ or an IP-ADV dominated by a PP-CONJ projected by a conjunctive particle, i.e., (PP-CONJ (IP-ADV ...) (P ...)). See section 2.6.3 for discussion of the particles. When there are multiple conjuncts, this arrangement gives a stacked list effect, so that, for example, clause1 is dominated by clause2 which is dominated by clause3. Consult section 5.4.2 for detailed explanations.
Distinguishing between instances of IP-ADV-SCON(-CND) and IP-ADV-CONJ will end up in vain if no criterion underlies it.
Here are our tentative criteria: For clauses ending either with predicates in the adverbial inflection (ren-you-kee, 連用形) or with the particle て, either -CONJ or -SCON(-CND) is assigned to each use listed on pp. 282–287 of altcite:[Nihongo08-2008-en]. The conjunctive(並列), contrastive (対比) and continuative (継起) uses are IP-ADV-CONJ instances, the conditional (順接条件) use is translated into IP-ADV-SCON-CND, and the other uses are IP-ADV-SCON instances.
The nature of clauses ending with a conjunctional particle or adverb is determined by the very particle or the adverb. Below is a selective list:
The particles of exclusitivity だけ and ばかり, and also the degree particle くらい can be adjoined to a finite clause, which is labelled IP-ADV.
やれる だけ やっ て み よう 。
だが 、 それ を 見届ける だけ の 生命 が 私 に ある かどうか の 自信 は なかっ た 。
どの 演奏 も 若さ と 情熱 が あふれん ばかり に 満ち て い た 。
ともかく 、 両足 を ふだん と は ちがう くらい 高く 上げ た 。
IP-ADV can be also found in adnominal て-clauses followed by の.
生まれつい て の 芸術家
見 て の とおり 、 道 から 部屋 の 中 を 覗きこめる ん だ 。
CP clauses are those projected by sentence-final particles (including quotation markers) which are separated from the rest of the sentences, which in turn become their inner complements. A CP clause contains:
Instances of CP are classified into CP-QUE (questions), CP-EXL (exclamatives), CP-IMP (imperatives), CP-FINAL (projections for sentence final particle), and CP-THT (complementizer clasese). This section deals with the first four types of CP. For CP-THT, see section 6.18.
Clauses describing the content of questions (IP-SUB, FRAG, etc.) are placed directly under a CP-QUE along with a question particle (P-FINAL) if any. Generally, question particles accompany question clauses, but not in some cases.
Questions are categorized into direct and indirect ones, which we do not differentiate by means of labels. Each case is elaborated below.
Clauses describing the content of direct questions (IP-SUB, FRAG, etc.) are put under CP-QUE projections.
目ざまし が 鳴ら なかっ た の だろう か 。
あなた 、 明日 の 会議 お出 に なる 。
「 それから ?
頭痛 が ひどい って ?
One context for an indirect question is where a question phrase (CP-QUE) is followed by a complementizer (P-COMP), the whole forming a complement clause (CP-THT).
最後 に 自分 に 神 を 信仰 する か と 尋ね た 。
There are also instances where an indirect question is formed without an accompanying complementizer particle. In these cases the question phrase (CP-QUE) is placed directly under the superordinate IP, with an appropriate extension tag on the label. Regarding the extensions for CP-QUE in indirect questions, see section 6.17.
新聞紙面 の 中 で 、 スポーツ面 が どれ だけ 、 復興 の 力 に なる の か 、 正直 分から ない 。
Moreover, we take two analyses for the particle か appearing in coordinated instances of CP-QUE, i.e., P-FINAL and P-CONN. See section 3.5.4 for the difference of these two かs.
Exclamatives are tagged CP-EXL. The constructions we treat as exclamatives at present in this corpus are:
なんて 奇妙 な 状況 な の だろう !
「 ああ 、 なんと 彼 を 愛し て おら れ た こと か 」 。
Note that such sentences that consist only of a noun phrase with an exclamation mark are not labelled CP-EXL.
「 うわぁ 、 おいし そう な ケーキ ! 」
こんな に うれしい こと が ある なんて !
Imperatives are labelled CP-IMP. Instances of CP-IMP with verbs in imperative or te-form as main predicate directly dominate an IP-SUB, while other instances of CP-IMP can dominate a FRAG. At present, items of the types listed below are treated as imperatives.
「 こっち に 来い よ 。 」
これ に お名前 を ご記入 の 上 ご提出 ください
投票箱 閉鎖 。
取り替え て ください 。
「 見 て !
こら 、 笑う な 。
/ 福 が いっぱい あり ます ように !
さっさと 行っ た 。
( 戒名 と 道号 について は 「 戒名 」 の 項目 を 参照 の こと 。 )
Subjects of imperatives are explicitly annotated. In cases where they are not expressed, a null pronoun (NP-SBJ *hearer*) must be supplemented.
「 助け て !
If an NP referring to the hearer is marked with a role particle or toritate-particle, it is regarded as the subject and an extension -SBJ is added to the particle phrase.
君 が 行っ て くれ 。
君 は ここ に い なさい 。
If there is a bare NP corresponding to the hearer, frequently it is marked as an NP-VOC and associated with a null subject in the following way: (i) Sort information is added to the NP-VOC. (ii) A following null *hearer* pronoun that shares sort information with the NP-VOC is introduced. See section 3.2.1.2 for examples of the treatment for NP-VOC.
田中君 、 取引先 に 資料 を 送り なさい 。
There are cases where a bare NP refers to an addressee of a command but cannot be regarded as a vocative in any sense. In such cases it is marked as NP-SBJ.
◎ 前向き な 姿勢 、 政府 示せ
When clauses accompanied by sentence-final particles (P-FINAL) do not represent speech acts that fall under the categories of questions, imperatives, or exclamations, they are assigned the label ‘CP-FINAL’. The sentence-final particle is placed directly under this node, sister to a preceding IP-SUB. Below is an example with sentence final particle よ (See section 2.8 for additional details).
私たち は 船 を 着陸 さ せ なけれ ば なら ない の だ よ 。 」
A sentence-final particle can follow a fragment (FRAG) too, as in the following example:
行き たい ん だ けど なー 。
If a constituent corresponding to a position in a preceding clause is postposed after a sentence-final particle, it is placed sister to that particle. Such expressions are treated as discontinuous constructions (see section 1.4 for details on these).
「 文字 だ よ 、 確か に 。 」
Right dislocated constituents are analyzed in the same way also in cases where there is no sentence-final particle available. The constituents and the original clauses are put together under an exocentric CP.
世界 を 動かし ます それ !
Most instances of CP take an IP as their complements. In other words, an IP occasionaly has CP expansions. A crucial difference between an IP and a CP is that the former always exists wherever there is a predicate while this is not the case for the latter. A CP layer is projected only when:
Otherwise it never appears.
All instances of IP that are dominated by CP layers are instances of IP-SUB (quasi-matrix clauses) (section 5.1.1).
Some CP clauses are embedded in other clauses as either arguments or adjuncts. The outermost layer (but not inner ones) of an embedded CP must be marked with syntactic role extensions such as -OB1 and -ADV (section 1.6.3).
「 大丈夫 ? 先生 ! 」 と 驚い た 声 で 僕ら は 聞い た 。
In regard to control environments (section 5.3) and ATB extraction (section 5.4), CP layers have no direct effects upon them. The only thing to note here is that, an IP-SUB layer, which is often the CP complement, necessarily blocks them.
A null subject whose antecedent is bound to be some argument of superordinating clauses can be found in adverbial subordinate clauses (IP-ADV-SCON) (occasionally with a further -CND extension), in gapless adnominal clauses (IP-EMB), in nomianlized clauses (IP-NMZ), and in small clauses (IP-SMC). Such an anaphoric relation is called ‘control’ in linguistic schools. The null subject is commonly referred to as ‘(big) PRO’. However, due to its conflict with our annotational labels for overt pronominals, we will instead use the term ‘controllee’ hereafter.
There is no explicit annotation of nodes corresponding to controllees using *pro* or the like. Instead, the downstairs clause is left without any node position at all. This practice follows a general ambition (e.g., seen with the annotation of unindexed null elements in section 1.5) to avoid indexing annotation whenever possible. Control relationships can be read directly from the structures assigned by the annotator. This section describes the annotation configurations that create control environments. Table 5.2 indicates whether or not a clause type allows a control environment.
Table 5.2: Control environments
| Syntactic tag & Control environment |
|---|
| IP-MAT (matrix clause) & no |
| IP-SUB (clause under CP layer) & no |
| IP-REL (relative clause) & no |
| IP-ADV-CONJ (coordinating conjunctive) & no |
| IP-ADV(2)-SCON(-CND) (subordinate conjunctive / conditional) & yes |
| IP-EMB (gapless adnominal clause) & yes |
| IP-NMZ (nominalized clause) & yes |
| IP-SMC (small clause) & yes (obligatorily) |
As Table 5.2 shows, control relations are not established across IP-REL (relative) or IP-MAT (matrix) layers. Nor are they allowed to straddle IP-SUB clause layers, which obligatorily call for a dominating CP. A caveat is that the opposite does not hold; a CP-THT can contain a small clause (IP-SMC) and yet does not hinder control relations that come from IP-SMC and go beyond it. As for adverbial clauses IP-ADV(2)-SCON(-CND) (and those instance of IP-ADV that project a PP), and coordinating conjunctive clauses IP-ADV-CONJ only the former license control relations.
A controllee may entertain multiple candidates, which are arguments of the immediately superordinating clause, for its antecedent. A particular accessibility hierarchy referring to grammatical roles will come to decide the actual antecedent. The hierarchy is as follows (the left being more prioritized):
There is one exception to the accessibility hierarchy. Some subordinate clauses specify for an antecedent with a subject grammatical role. When the accessibility hierarchy needs to be disregarded to accomodate this behaviour, the clause is marked as IP-ADV2 possibly with the further extensions of -SCON. Clauses with conjunctional particle ながら are good examples of this.
More generally, with all control environments (see Table 5.2 above), it is possible to ensure dedicated control by the subject of the containing clause by adding a ‘2’ to the clause tag of the control environment (IP-EMB2, IP-SMC2, and IP-NMZ2).
For antecedents with grammatical roles other than subject, precedence of the antecedent with respect to the controllee is a condition on accessibility for most control environments. On the other hand, a subject role antecedent can follow an adverbial clause (IP-ADV-SCON(-CND)). Furthermore, with respect to small clauses (IP-SMC) and clauses whose subject are obligatorily controlled by matrix subjects (IP-*2), the precedence condition does not apply for antecedents of any role.
The following sections demonstrate placement patterns for the subordinate clauses that do allow control, and the relations of control that arise.
The present section concerns instances of IP-ADV-SCON(-CND). Constructions where an IP-ADV is the complement of a PP-SCON(-CND) are subsumed in this discussion. When IP-ADV is disambiguated with -CONJ (see section 5.4), anaphoric relationships between upstairs anaphors and null elements do not fall under the description of control.
For instances of IP-ADV, if there is no -SBJ local to the adverbial clause, then a controllee can inherit its reference from a controller in nearest higher clause layer, favouring -OB2 over -OB1 over -SBJ2 over -SBJ when any of these is present and provided such noun phrases are accessible. To be accessible a controller should precede the IP-ADV, with the exception of the subject which is always accessible.
Figure 5.1 illustrates the scenario of an -OB2 acting as a controller.
Example (5.50) has an -OB2 which controls the subject of an IP-ADV under a PP-SCON.
ランドセル は 孫 に , 小学校 に 上がっ た ので あげ まし た 。
Figure 5.2 illustrates a scenario in which an -OB1 controls the subject of an IP-ADV.
The following are examples with -OB1 being controllers into IP-ADV under a PP-SCON or a PP-SCON-CND:
その お菓子 は , まずかっ た ので 弟 に やっ た 。
ビール は よく 冷え て い て も 飲み たく ない 。
Figures 5.3 and 5.4 illustrate scenarios of -SBJ acting as controller into IP-ADV-SCON(-CND) and IP-ADV under PP-SCON(-CND).
Figure 5.3: SBJ as controller into IP-ADV-SCON(-CND)
Figure 5.4: SBJ as controller into IP-ADV under PP-SCON(-CND)
The following are examples of -SBJ at various positions controlling the subjects of IP-ADV-SCON and IP-ADV under PP-SCON.
寂しかっ た ので 私 は 友人 を 呼ん だ 。
ギョッと し て 武士 は 足 を 早める 。
Control into IP-EMB occurs as with IP-ADV (see section 5.3.1) with a favouring of:
when any of these is present. In the same way as IP-ADV, the position of controllers, as well as the accessibility hierarchy, is taken into consideration. That is, a controller must precede the IP-EMB whose subject it controls, with the exception of subject controllers, which are always accessible to controllees regardless of their locations.
The following illustrates the scenario of an -OB2 acting as a controller into IP-EMB.
Figure 5.5: OB2 as controller into IP-EMB
An example of an -OB2 as a controller:
手紙 は 太郎 に ここ へ 来 た とき に 渡し た 。
The following illustrates the scenario of an -OB1 acting as a controller into an IP-EMB.
Figure 5.6: OB1 as controller into IP-EMB
An example of an -OB1 as a controller:
二郎 は たこ焼 を 熱い うち に 太郎 に 渡し た 。
The following illustrates the scenario of a -SBJ acting as controller into an IP-EMB.
Figure 5.7: SBJ as controller into IP-EMB
Figure 5.8: SBJ as controller into IP-EMB
Control from the subject of the containing clause can pass through an NP argument layer into an IP-EMB complement provided the NP restriction contains in addition to the IP-EMB complement only a head noun, annotated N.
The following are examples of -SBJ as controllers into IP-EMB:
試験 を 受ける 前 に 、 トイレ に 行っ た 。
小さかっ た ころ 私 は 犬 を 怖がっ て い た 。
Control into an IP-NMZ parallels those into IP-ADV-SCON(-CND) and IP-EMB with a favouring of:
OB2 > OB1 > SBJ2 > SBJAs with IP-ADV-SCON(-CND) and IP-EMB, a controller should precede the IP-NMZ in order to be accessible to the controllee, with the exception of the subject, which is always accessible regardless of its position.
The following are examples of -SBJ as controller into IP-NMZ:
彼 の 子孫 は それ を 所有 する に いたる であろう 。
辺 の 部分 に フラップ を 作る に は 、 円 の 2分の1 が 必要 です
Incidentally, no example is attested at the moment where arguments other than subjects control into IP-NMZ.
IP-SMC is never annotated as containing a subject argument. Consequently it is an obligatory control environment. Typically forming a part of a complement of the predicate of the immediately containing clause, IP-SMC has its control relationship determined from among the full range of arguments present in the containing clause. This favours -OB2 over -OB1 over -SBJ to act as the controller when any of these are present and irrespective of where these noun phrases occur.
The following figures illustrate an -OB2 controller which obligatorily binds the subject of an IP-SMC.
Figure 5.9: OB2 controller into IP-SMC
Figure 5.10: OB2 controller into IP-SMC
The figures below are about -OB1 binding IP-SMC subjects.
Figure 5.11: OB1 controller into IP-SMC
Figure 5.12: OB1 controller into IP-SMC
Examples with -OB1 as controllers:
私 は 弟 に 買い物 に 行っ て もらっ た 。
わたくし の 考え を 述べ させ て いただき ます 。
The following figures illustrate cases where a -SBJ controls the subject of an IP-SMC.
Figure 5.13: SBJ controller intro IP-SMC
An example in which a SBJ is a controller:
死に たく 思う 。
Those IP instances which are additionally marked by 2, namely IP-ADV2-SCON(-CND), IP-EMB2, IP-NMZ2, and IP-SMC2, must be under control relations with an external argument, as with IP-SMC. What differs from IP-SMC is that the controller, as well as the controllee, must be a subject (-SBJ or -SBJ2). The position of subject antecedents does not matter.
The following is an example of a -SBJ as controller into an IP-ADV2-SCON:
絵本 を 買っ た 子供 が 、 それ を おやつ を 食べ ながら 読ん で い た 。
When an upstairs candidate for antecedence, which is uniquelly determined by the subcategory of the embedded IP and the accessibility hierarchy, is not what is meant by the sentence, control must be manually disabled by introducing a null NP-SBJ (*pro*, *speaker*, *hearer*, *arb*, etc.), whichever is appropriate (see section 1.5), into the controllee position. This is because control relations emerge whenever they are licensed unless otherwise specified.
In example (5.66), *arb* is internal to the IP-EMB to prevent the matrix SBJ controlling an empty subject position in the IP-EMB.
私 の 趣味 は 料理 を する こと です 。
There are cases (e.g., involving weather predicates), where neither control nor ATB antecedence (section 5.4) should be allowed. Prevention of all anaphoric relation is achieved by introducing (NP-SBJ *exp*) into the controllee position, as in example (5.67).
春子 は 寒く なる と 学校 に 来 なく なる 。
When there is a co-reference between an upstairs argument and a null subject position in a control environment, but control is inappropriate due to either (i) the correct antecedent is not accessible to the null position, or (ii) the null position not being a subject position (which, as has been being explained in this section, is out of the scope of the control mechanism), or (iii) both (i) and (ii), an antecedent relationship can be established through the explicit annotation of a zero pronoun (NP-SBJ *pro*) plus sort information shared between the zero pronoun and the antecedent.
ダウンロード し て 印刷 すれ ば 、 学校 や 家庭 で 手軽 に 取り組める 。
As is mentioned in sections 5.1.5.1 and 5.1.5.2, IP-ADV-CONJ (or IP-ADV under PP-CONJ) are (the only) instances that allow Across the Board (ATB) extraction. All constituents that precede and are sister to the second-to-last conjunct (that is, the last (IP-ADV-CONJ ...)) are ‘inherited’ by each conjunct. section 5.4.1 considers cases with only two conjuncts, while section 5.4.2 illustrates the more than two conjuncts scenario.
As with control, a zero pronoun that takes an ATB extracted antecedent is not explicitly annotated. But also note that, unlike control, any constituent can be inherited by coordinate conjuncts, provided that it is (ATB) left-posed. Furthermore, the grammatical role of an extracted constituent is preserved in each conjunct (another contrast with control).
In the example and figure below, both conjuncts are saturated. Nothing holds an antecedent relation with a null element in IP-ADV:
また 一方 で 、 地元マスメディア も 地元 の 人材活用 を する よう に なっ て 、 バンド ・ タレント ・ 芸能人 ・ モデル など の ローカルタレント が 増加 し て 、 隣県 の メディア に 進出 する 者 も 表れ た 。
Figure 5.14: Saturated conjuncts
In the example and figure below, only the left-posed NP-SBJ holds an antecedent relation with a corresponding null element in IP-ADV.
それ は 1万5千 円 の もの です が , 今 は バーゲン で 1万 円 です 。
Figure 5.15: SBJ into conjunction
The following example is the same as above insofar as SBJ holds an antecedent relation, while the OB1 position is overtly filled in both conjuncts. But note that the adjunct 昨日は is additionally inherited by both:
昨日 は 刺身 を 食べ て , お腹 を 壊し て しまい まし た 。
In the example and the figure below, the left-posed SBJ and OB1 hold an antecedent relation with corresponding null elements in IP-ADV:
晩ごはん を 作っ て , 食べ ます 。
Figure 5.16: SBJ and OB1 into conjunction
In the figure below, the left-posed NP-SBJ, NP-OB2 and NP-OB1 hold antecedent relations with corresponding null elements in IP-ADV.
Figure 5.17: SBJ, OB1 and OB2 into conjunction
If there is an overt -CONJ then it is treated as a constituent of the superordinate IP, that is, an IP level word.
私 は 長い 間 この 動き を 観察 し て そして 参加 し て き まし た
As is explained in section 5.1.5.2, clause-level coordination is represented with the stacking of clause content so that the first conjunct's clause content is dominated by the second conjunct's clause content, which is dominated by the third conjunct's clause content, and so on. Clauses so stacked should be leftmost with regards to the content of the next higher clause, except when the next higher clause is the last in the stack, that is, the last conjunct. Also, with the exception of the highest clause, all other clauses in the stack are labelled as IP-ADV nodes that contain a -CONJ extension. By contrast, the final clause in the stack has the clause label that is appropriate for the conjunction of all the clauses (e.g., IP-MAT, IP-SUB, or IP-ADV). This is shown by examples (5.74) and (5.75), and the figure below.
暗く て 、 たくましく て 、 ちょっと もの悲しい リアス式海岸 は 、 僕 にとって の 東北 の 原点 。
僕 は 集め て 集め て 集め まくっ た 。
Figure 5.18: Stacked conjunction
When it comes to interpreting annotation that conforms to the stacked structure of Figure 5.18, we can think of the derived CONJP structure of Figure 5.19 with IML (clause intermediate level) nodes as constituting the basis for interpretation. Such derived structure (obtained with an automatic process) is essentially identical to the explicitly annotated CONJP structure seen with non-clausal coordination in section 3.5.
Figure 5.19: Interpretation with IML
This chapter describes the annotation of different constructions.
There is a common sentence type in Japanese called a ‘double subject sentence’:
神戸 は 夜景 が きれい だ 。
In the example above, both は and が mark NP layers that are annotated as grammatical subjects. The second subject (closest to the predicate) is interpreted as having a simple subject role with regard to the predicate, independent of other factors. The first (leftmost) subject is dependent on a relationship with either the second subject, or with the predicate, or with the predication that these two form (discussed in detail below). According to our annotation framework, the first subject (regardless of its particle marking) is tagged as -SBJ, and the second が-marked subject is assigned the label -SBJ2. As arguments, either can function as a controller in a superordinate clause.
In non-relative clauses, a non-focus first subject is typically marked with a toritate particle such as は, も, なら, だって, etc. A further characteristic of this construction is that the first subject can be marked with が given the right context, either in a relative clause
or with some special focus.
Sentences of this kind are subdivided into at least three types depending on the relationships that obtain between NP-SBJ, NP-SBJ2, and the predicate. (see altcite:[Nihongo-2009-5-en], pp. 186-189). (The distinction between these types is not specified in the annotation, but an awareness of the types is useful for recognising the general construction.)
In the first type of double subject sentence, the first は-marked subject can be construed as the nominal complement of the second が-marked subject. Therefore, a sentence ‘A は B が C (だ)' can be converted into ‘A の B が C (だ)' without essentially changing the meaning. For example,
The second type is formed on a nominal predicate, where the first subject can be construed as the complement of the noun that occurs as predicate. In other words, ‘A は B が C だ’ is almost equal to ‘B が A の C だ’. Typically the head of the nominal predicate denotes some crucial or integral aspect of NP-SBJ in these sentences. Here are examples:
In the third type, the first subject does not stand in any particular grammatical relationship to the second subject or to the predicate individually, but has a relationship to predication that the pair forms. For example,
In this sentence, このにおいは has no relationship to ガス or to 漏れている in terms of complement relation or grammatical role, but has a summational relation to the predication they form.
The double subject construction is so robust in Japanese that a null NP-SBJ is construable when it is relativised, or controlled by an NP in a superordinate clause or retrievable from context. In the example below, The speaker is NP-SBJ of the IP-MAT, while at the same time functioning as controller into the IP-ADV, which has two subjects.
実家 が 塩釜 の 海産物屋 な ので 、 浜 の におい の 中 で 育っ た 。
A careful distinction must be made between this kind of sentence and sentences with predicates that call for a が-marked NP-OB1. As explained in section 4.7 in some detail, in constructions with が-marked NP-OB1, the NP-OB1 can be left unexpressed and the NP-SBJ is still interpretable as a subject of the predicate. In most double subject sentences, however, the NP-SBJ is not simply interpretable as a subject of the predicate. Furthermore, only a limited number of predicates subcategorise for a が-marked object, whereas there is hardly any restriction on the type of predicate in a double subject construction, provided that the relevance of the NP-SBJ can be drawn.
Note that there are constructions of the type exemplified below which do not fit into any of the categories listed above and seem to be derived from topicalisation and left-shifting of the first NP plus exhaustive focus marking on the second NP:
It is not clear whether these should be treated as double subject sentences.
It is also a well known fact that triple and quadruple subject sentences are possible: 日本は女性は平均寿命が長い。 The question of how many instances of NP may be assigned subject status in one clause is another question that needs to be addressed. In this corpus, there is no other extension tag for subjects than -SBJ and -SBJ2.
Also note that the definition of the first type of double subject sentence is satisfied by constructions with the so-called ‘transitive (possessive) verb ある’, but our annotation treats the second が-marked argument for these constructions as NP-OB1.
Infrequently there are examples in which a head noun appears to have been deleted and a の-headed PP functions as a restrictor on some generalized entity:
すごい いい デザイン の が あっ て これ に 決め た !!
This phenomenon is called ‘N-bar deletion’. Note that the head is left empty; that is, with no overt annotation whatsoever.
Infrequently there are examples of coordination in which a head predicate appears to have been deleted from a non-initial conjunct:
右 から 開く と 防災マップ 、 左 から 開く と 高齢者見守り の マップ と なる よう に デザイン し た 。
This phenomenon is called ‘Right-node raising’. Note that the head of the first conjunct is left empty; that is, with no overt annotation whatsoever.
最終的 に 舞台 の アンネ役 は スーザン・ストラスバーグ が 、 映画 の アンネ役 は ミリー・パーキンス が 演じ た 。
太郎 は 小説 が 、 花子 は 詩 が 、 好き だ 。
Note that in this construction the first conjunct without a predicate forms an IP-ADV-CONJ.
There are structures involving parallelism that are marked up in ways analogous to right-node raising. In the example below paralleism is expressed by conjunction of IP instances where more than one node is ‘left behind’ by the first conjunct. These are indicated by headless categories within the IP conjunct, in addition to the headless IP itself.
参加者 の アンケート に は 、 過去 に 経験 の ある 人 から 「 メンバー が 変わる と 答え も 変わっ た 」 、 初めて の 人 から 「 こんな に 多様 な 意見 が ある と は 思わ なかっ た 」 など の 回答 が 寄せ られ た 。
There is a class of costructions called ‘depictive (absolute) clauses’ that are similar to right-node raising constructions in lacking a verb. However, these clauses are unique in that they are adverbial to a following predication, rather than being parallel to it. Generally, they are of the form [A を B に] and placed under IP-ADV-SCON, unlike right-node raising constructions. They are further divided into two types with two different structural descriptions.
If に can be replaced with として without changing the meaning of the sentence, then [B に] is analyzed as a small clause (IP-SMC-OB1), with に as a copula (AX) in infinitive form:
__インターネット を介して 音楽 や 映画 を 配信 できる よう に なっ た 。
親鸞 は 、 ここ を 拠点 に 精力的 な 布教活動 を 行う 。
If に cannot be replaced with として, then [B に] is analyzed as a post-positional phrase (PP-OB2), with に as a particle performing a peripheral grammatical role:
オリンピック を 前 に 、 選手たち は 泳ぎ込ん だ .
地図 を 手 に 目的地 を 探し た 。
If there is concatenation of IP-MAT, IP-IMP, or clauses of a CP level within a quotation, or followed by a particle that modifies a noun with a content complement (という, といった, etc.), then the concatenated elements are placed directly under an instance of the multi-sentence tag. FRAG and INTJP, when forming distinct utterances, can also be placed directly under multi-sentence.
「 イライラ が 募る の か 。 明らか に 震災後 の 現象 だ 」 と 校長 は 対応 に 困惑 する 。
私的日記 に とどまら ない 、 「 誰 か の ため に なる 」 「 役 に 立つ 」 「 琴線 に 触れる 」 よう な ブログ が 、 読み手 の 共感 を 得る 。
Note that the multi-sentence tag by itself does not create a syntactic island such as IP-SUB creates under CP-THT. Rather, it is the practice of allowing only utterance-level categories (namely, IP-MAT, IP-IMP, FRAG, INTJP, CP-QUE, CP-EXL, CP-FINAL) to appear under multi-sentence that produces the island effect.
Focused pseudocleft constructions are analysed as copular clauses, typically with -SBJ comprised of (N の) as the head noun, modified by an IP-REL with a trace. A subject so comprised forms the presuppositional part of a focused pseudocleft sentence. The focus part of the sentence is a copular predication, typically formed on an NP-PRD.
太郎 が 勉強 し て いる の は 日本史 だ 。
The focus part may be occupied by the projection of a particle phrase (PP-PRD).
「 わたしたち が 信じる の は 、 もう あなた が 話し て くれ た から で は ない 。
On the other hand, there are pseudoclefts which are annotated as involving not an IP-REL, but an IP-EMB. A typical case involves a focused element which expresses a reason for the presuppositional part. For such cases, it is not clear what kind of trace should be placed if an IP-REL were assumed.
あなたがた が 信じ ない の は 、 わたし の 羊 で ない から で ある 。
Furthermore, conjoined pseudoclefts are analyzed as having an IP-EMB uniformly if one of them must be so analyzed, even when an IP-REL analysis is possible for another conjunct.
二人 が 知り合っ た の は 友人 を介して で あり 、 ヘプバーン と ドッティ と の 結婚生活 が 終わり を 迎え よう と し て い た 時期 だっ た 。
Note that there are copular clauses that have the same surface structure as focused pseudocleft constructions, but don't involve the concealed question/answer pairing that is characteristic of focused pseudocleft constructions. There is no special annotation to distinguish this construction from pseudoclefts.
津波 に 対抗 し て 堤防 を 高く する の は 、 技術力 に 依存 し た 20 世紀 型 の 発想 。
Double-ヲ constructions, in which a transitive verbal noun taking an を-marked PP-OB1 is itself followed by を and VB0, occasionally occur in formal speech situations (e.g., in 国会会議録). In these instances, the second を is treated as a clausal constituent P particle directly dominated by IP:
ただいま 議題 と なり まし た 法律案 につきまして 、 経済産業委員会 における 審査 の 経過 と 結果 を 御報告 を 申し上げ ます 。
また 材料 も 厳選 さ れ た もの を 素材 に応じて あくぬき など を し ながら 、 味 を ひきだす 技術 が 要求 さ れる 。
Direct passive sentences are those in which the -SBJ has the semantic role of Patient or Recipient with respect to the core predicate, and the NP with the semantic role of Agent with respect to the core predicate is, if any, marked with a role particle (に, から, によって, etc.). One way to think of this is that an underlying -OB1 or -OB2 has been promoted to subject position, while an underlying subject has been demoted to an oblique position. The NP or PP with the Agent role is given the extension tag -LGS (a logical subject). The passive auxiliary verb is labelled PASS.
そう すれ ば お父さん から ぶた れ ない だろう 」
Passive constructions can feed the formation of double subject sentences:
ジョン は 論文 が 受理 さ れ た 。
金庫 の 中身 は 、 大部分 が テーブル の 上 に 出さ れ て い た 。
In the passive construction with ditransitive verbs, either -OB1 or -OB2 may be promoted to subject position. The remaining object does not change its grammatical role with respect to the core verb.
二 第三者 に 提供 ___さ れる 個人データ の 項目
This entails that the only time -OB1 and PASS co-occur in the same IP is when the subject is promoted from an -OB2 position.
難しい 時代 に 直面 する 今 、 私たち は 構想力 を 問わ れ て いる 。
When the logical subject appears in the same clause as an -OB2, it is usually marked with a particle other than に (e.g., から, によって, により).
その 後 、 禅宗 の 最高峰 を 極め た 臨済宗 は 、 南宋時代 の 中国 に 渡り 学ん だ 栄西ら によって 、 鎌倉時代 に 日本 に 伝え られ て いる 。
Causative constructions are valence increasing constructions in that an additional argument (viz., the ‘causer’) is added to the argument frame of a predicate. In a causative construction, the core predicate (VB), its non-subject (internal) arguments, semantic Agent of the core predicate (i.e., the ‘causee’, typically marked with either に or を) and syntactic subject (the ‘causer’) all have sister positions under IP. The semantic Agent is tagged with the extension -CZZ. The auxiliary verb (VB2 (さ)せる) immediately follows the core predicate.
私 は 弟 に ケーキ を 食べ させ た 。
「 どこ から パン を 買っ て き て 、 この 人々 に 食べ させ よう か 」 。
In a causative passive construction, an -SBJ-marked element of the local clause is the syntactic subject of the entire passivized predicate, (combined with both the auxiliary verb (さ)せ (VB2) and the auxilliary verb られ (PASS), and also the logical Agent of the core verb. A local -LGS argument is the logical Agent of the causation event expressed by the causative morpheme (さ)せ (VB2) attached to the core predicate.
高津さん は その 事件 を 調べ させ られ た 。
そこで 妹 は 母親 の 忠告 によって 自分 の 決心 を ひるがえさ せ られ たり し て は い なかっ た 。
In contrast to the direct passive (see section 6.9), in an indirect passive construction the auxiliary verb (ら)れ is assigned the label PASS2. Indirect passive constructions are valence increasing constructions in that an Affected -SBJ argument is added to the argument frame, while the logical subject of the core predicate is demoted to -LGS and marked with に. These two arguments are sisters of the the core predicate (VB) combined with the passive morpheme (PASS2 (ら)れ), together with any internal arguments, which are marked for roles with respect to the core predicate.
太郎 が 花子 に 泣か れ た 。
太郎 は 先生 に 絵 を ほめ られ た 。
テアル constructions are stative constructions that are primarily used to background agentivity in describing the results of volitional action.
The テアル construction is divided into two analyses according to the transitivity of the core predicate. When a combining stem of a transitive verb (or verb compound) is immediately followed by (P-CONN て) which in turn is followed by some form of (VB2 ある), if the underlying object of the transitive verb appears as a が-marked NP (or as a の-marked NP in a noun-modifying clause), that NP is labelled NP-SBJ and a node (PASS *) is introduced immediately after (P-CONN て). Furthermore, in cases where a CP-THT is used as argument of a verb of communication in the てある form (e.g., 書いてある), without any overt subject in the same clause, the CP-THT is extended with -SBJ (see section 6.18.2) and (PASS *) is inserted in the same way. In this treatment, the underlying subject is completely suppressed, and the underlying object is promoted to subject, but other grammatical roles (specifically, -OB2) remain unaffected.
子供 が 笑っ て いる 写真 が 置い て あっ た 。
それから 鼠花火 という の は 一 つ ずつ 輪 に なっ て い て 箱 に 詰め て ある 。
For this construction, if the underlying object of the transitive verb appears as an を-marked NP, a bare NP, or an NP marked with a toritate particle, etc., that object is treated as -OB1. When no subject seems to be involved, a (NP-SBJ *exp*) subject is introduced. As the argument structure and the case marking are in their usual alignment, no node (PASS *) is introduced.
丁度 便所 の 坑 の 傍 に 、 実 を むしり 残し た 向日葵 の 茎 を 二三 本 縛り寄せ た の を 、 一 本 の 棒 に 結び附け て ある 。
この 生け花 は なかなか 見事 に 生け て ある 。
The treatment for both cases is extended to traces in IP-REL and for null elements in general.
それから 水 に 漬け て ある 豆 だとか 慈姑 だとか 。
When a combining stem of an intransitive verb (or verb compound) is immediately followed by (P-CONN て) which in turn is followed by some form of (VB2 ある), an NP subject of that verb is labelled NP-SBJ and no node (PASS *) is inserted. Examples can involve either unaccusative verbs,
台所 に 、 まだ 酒 が 残っ て 在る 筈 だ 。
黄色く 薄濁り し た 液体 が 一ぱい つまっ て 在る 一升瓶 は
or passivized verbs.
汚れ た すじ が 四方 の 壁 に沿って 引か れ て ある し 、 そこかしこ に は ごみ と 汚れもの と の かたまり が 横たわっ て いる 始末 だ 。
Clausal conjunction is seen with the テアル construction as well:
瓦斯煖炉 が 焚い て 、 電燈 が 附け て ある 。
てもらう, てあげる, and てくれる belong to one and the same group of secondary verbs in the traditional study of Japanese. In our annotation, however, てもらう and てあげる/くれる are treated differently. Even in its subsidiary usage attached to a main verb, てもらう is labelled VB and embeds an IP-SMC (see section 5.1.4). See section 2.3.3 for control over IP-SMC by a に-marked PP.
私 は 父 に カメラ を 買っ て もらい まし た 。
いただく, the humilific form for もらう, undergoes the same analysis.
By contrast, あげる and くれる are tagged as VB2 and do not embed the predicate and accompanying consitituents they are attached to. All the constituents are placed at the same level directly dominated by IP-MAT.
私 は 友達 に お金 を 貸し て あげ た 。
道男 は 私 を なぐさめ て くれ まし た 。
The same applied to さしあげる and くださる, the honorific forms for あげる and くれる.
In a comparison of two activities or events, typically the one forming the basis of comparison will be expressed in a finite clause marked directly with particle より, while the event that is characterised is typically an -SBJ formed on a formal noun:
理屈 を ひねくり回す より 、 やっ て みる こと が 大切 だ .
Verbs denoting motion are occasionally modified by either activity noun phrases or infinitive clauses indicating the purpose of motion. For noun phrases, an activity noun (e.g., 買い物) --including verbal nouns like 遊び-- heads a phrase to which role particle に is attached. For infinitive clauses an activity verb in the infinitive inflection heads an IP-NMZ to which role particle に is attached. The role particle に heads a PP with the extension -PRP.
彼女 の お母さん は 買い物 に 行き まし た 。
メロス は 、 それゆえ 、 花嫁 の 衣裳 やら 祝宴 の 御馳走 やら を 買い に 、 はるばる 市 に やって来 た の だ 。
For infinitive clauses indicating purpose of motion, typically the subject of the verb of motion controls into the infinitive clause. However, non-subject arguments of caused-motion predications such as 行かせる and 派遣する can also serve as controllers for purpose of motion clauses.
小僧 ___に ___きき ___に ___こ させる だけ で ほんとう に 十分 で は ない だろう か
Note that purpose of motion phrases headed by activity nouns don't have argument structure.
A typical indirect interrogative contains a CP-QUE which is followed by a particle (P-ROLE) and serves as an argument of a predicate in the next higher clause. When an indirect interrogative has a core grammatical role, a corresponding extension tag is added to the PP which dominates it. Below are some examples.
その 日 何 を し て い た か を 詳しく 説明 し なさい 。
この 仕事 は 苦労 も 多い が 、 それだけ やりがい も ある 。
折り紙 の 最強 の ツール は 、 我々 が 部品 を どうやって 作る か に 関係 し て い ます
どちら が 県連 の 推薦 を 受ける か で もめ 、 木村太郎 衆院議員 が 県連会長 を 辞任 する 騒動 に まで 発展 し た 。
ここ で は , 企業 が いかに CSR に 介入 し , どのような 問題 に 関心 を 持っ て いる か について 討議 し たい と 思い ます 。
When a CP-QUE is followed by a toritate particle which can be replaced with a role particle for marking a core grammatical role without change of meaning, a suitable extention is added to the PP projected by the toritate particle.
鈴木さん が 大学 に 入っ た かどうか は 知り ませ ん 。
どうして 食っ て いる か さえ 分ら なかっ た 。
There are also cases where CP-QUE without a role or toritate particle can be naturally complemented with a role particle such as が or を. In such cases, an extension tag is directly added to CP-QUE. However, appositive interrogatives require further special treatment (see the next section).
自分 が 何 を 求め て いる の か 頭 の 中 で はっきり し た の です
山 は もう 寒い か 聞い て ください 。
In addition to the patterns discussed above, there are instances of utterance-level constituents that seem to be closely related to the contents of the arguments as well as to the meanings of the predicates, but are not necessarily adjacent to the arguments in question. In the case of CP-QUE specifically, an overt argument is headed by a relational noun that is interpreted as a concealed question. Despite the close relationship, there is no adjacency condition between the CP-QUE and the argument, so a structural relation indicating apposition is not sufficient to describe the pattern. In order to indicate this relation, a parenthetical expression PRN is introduced together with the argument, and the CP is related to that position through the use of *ICH* (Interpret Constituent Here).
In the following example, the interrogative clause 芸術界で今何が起こっているか can be interpreted as the direct object of the predicate 分かる because it is possible to insert the role particle が after the interrogative. At the same time, it can be regarded as having a semantic relationship to the following direct object 動向(が) . To annotate this structure, CP-QUE is put under PRN extended with some appropriate number (normally PRN-1). Then another PRN is added in the sister position of the N which heads the noun phrase related to the CP-QUE. This PRN has as its daughter *ICH* extended with the same number as that of the other PRN (normally *ICH*-1).
芸術界 で 今 何 が 起こっ て いる ___か 動向 が 分かる と さ れる もの でし た
The head of noun phrases associated with CP-QUE can be a pronoun or modified by a demonstrative determiner.
しかし 、 どうして いま 見える よう に なっ た の か 、 それ は 知り ませ ん 。
There are also cases where CP-QUE is not an argument of a predicate, but has an appositive relation to an argumental noun phrase. For example, when used with expressions such as 見当がつく, 意見が分かれる, and 議論をする, CP-QUE instances do not serve as arguments of these expressions, but describe the content of head nouns 見当, 意見, and 議論. In these cases, too, the tags *ICH* and PRN are used.
それに どっち に にげ て いっ たら いい の か けんとう も つき ませ ん 。
If a CP-QUE is not an argument of a predicate in the next higher clause and cannot be complemented with a role or toritate particle, the extension tag -ADV is added to it. Such adverbial interrogatives (CP-QUE-ADV) are mostly self-addressed.
その 勢い に 驚い た もの か 、 捕り手 は パッと 左右 へ 開い た 。
たとえば 、 紅茶 を 入れ た のに 、 どういう わけ か 、 コーヒー を 入れ た と 勘違い し て しまう 。
There are cases where it is not clear whether a question is adverbial or argumental. The example in (6.63) has an overt を-marked argument, but also contains a CP-QUE (which cannot be case-marked in this context). Although the CP-QUE might be allowed as a complement of the predicate under different circumstances, here it is marked as CP-QUE-ADV. Note that this adverbial interrogative is neither appositive to an N nor interpreted as self-addressed, unlike the examples above.
間違い が ない かどうか 、 書類 を よく 見直し て ください .
Interrogatives (CP-QUE) are sometimes used as a predicate followed by a copula (AX) such as だ and です. In this case, the extention tag -PRD is added to the label CP-QUE.
重要 な の は これら の シンボル が 何 を 意味 し て いる か です
Interrogatives (CP-QUE) are sometimes followed by a complementizer patricle such as と. This kind of ‘quoted’ interrogatives mostly serve as a complement of verbs of communication and cognition.
刑事 は 裏 の 窓 を 閉め た か と 聞い た 。
うん 、 行こ う かなー って 思っ てる 。
Furthermore, interrogatives can be followed by a complementizer particle such as という or との to form an adnominal clause (See also (3.30)).
問題 は 誰 が 社長 を やり 、 誰 が 社長代行 を やる か という こと だっ た 。
Typically, a CP-THT appears under an IP layer as complement of verbs of communication and cognition. In such cases, they are tagged with the extension -OB1. Such clauses do not actually have to be interpreted as direct quotations.
「 私 は むしろ あなた を 気に入っ てる の 」 と 彼女 は 言っ た 。
彼 が そんな に 美しい 死体 に なる と は 誰 も 思わ なかっ た でしょう 。 」
CP-THT instances can also, though infrequently, appear as argument with a role particle. In the following examples, CP-THT is followed by a role particle such as が and を. In such cases, the extension tag -SBJ or -OB1 is added to the PP projected by the role particle.
複数回答 の 結果 、 「 家計的 に 苦しい 児童 ・ 生徒 が 増え て いる 」 が 小中学校 とも に 最多 だっ た 。
ワニ君連 を 代表 し て 、 花束 を 持っ て お見舞い に 来 た アシ君 が 、 槇子 の ( 秋作さん 、 秋作さん ) を 聞い て しまっ て 、 これ を 『 社交室 』 へ 急報 し た 。
As the following examples show, a CP-THT followed by a complementizer particle such as って or なんて can be a subject without a role particle. In such cases, an extension tag is directly added to CP-THT.
漢字 を 覚える って 大変 です ね 。
そんな に いつ まで も しょげ て いる なんて 、 あなた らしく ない こと よ 。
In sentences in which a passivized verb of communication or cognition co-occurs with a CP-THT, the subject is often covert. In such cases, it is sometimes difficult to decide which element to regard as the subject. In the following example, the zero pronoun *speaker* is inserted as the subject which denotes the receiver of the content expressed by the complementizer clause.
8 時 まで に 来い と 言わ れ た ん だ 。
In contrast, in the following example, no concrete receiver of the communicated content is conceivable and thus it is difficult to regard it as having a receiver as the subject. Furthermore, there is no other noun phrase which can be though of as the subject. In such cases, the complementizer clause is regarded as the subject and the extension tag -SBJ is added to CP-THT.
運動 は 体 に いい と 言わ れ て いる 。
Also, in cases where a verb of communication or cognition is used with a CP-THT in てある constructions and no noun phrases can be regarded as the subject, the complementizer clause is annotated as the subject.
なんか 、 国際交流セミナー っていう やつ 、 ある って 書い て あっ た ん だ けどー 。
Instances of CP-THT, as well as CP-QUE, can be semantically related to both the predicate in the next higher clause and its argument noun phrase. In the following example, a CP-THT is the complement of the verb of cognition 聞きます and expresses the content of the -OB1 noun phrase 話(を) at the same time. To show such an appositive relation between a CP-THT and a noun phrase, *ICH*-n and PRN-n are used, as is the case with appositive interrogatives (CP-QUE).
やんちゃ で 済まさ れ ない こと が 多く 、 地域 の 方 から は 、 震災後 から 荒れる よう に なっ た と 話 を 聞き ます 。
The head of noun phrases associated with the complementizer clause can be a pronoun or modified by a demonstrative determiner.
坂本君 、 あのー 国際交流問題 とか 、 興味 ある とか 、 そういう こと 言っ て た からー
There are also cases where CP-THT is not an argument of a predicate, but has an appositive relation to an argumental noun phrase. In these cases, too, the tags *ICH*-n and PRN-n are used.
「 えい ! 」 と 初めて 声 を 掛け 、 右手寄り に ツツ――と 詰める 。
Sometimes, CP-THT instances do not serve as complements of the clausemate verb or show any semantic connection to other nouns in the clause. Among them, those which contain a volitional auxiliary such as (AX う・よう) or (NEG まい) and exress the purpose of the action denoted by the clausemate predicate are annotated with the extension -PRP (a special kind of -ADV):
僕 は 僕 を 子供 だ と ほのめかし た コター の じいさん に 腹 を 立て て い た けれども 、 彼 が 言い 終え なかっ た センテンス から 意味 を 引き出そ う と 頭 を 悩まし て い た 。
マラソン は 苦しかっ た が 、 友達 に 遅れ まい と 一所懸命 走っ た 。
CP-THT instances can appear under IP as adverbial elements when it is clear that they neither relate to their sister predicates or nouns, nor express the purpose of the action denoted by the sister predicates. They are marked as an adjunct with extension -ADV.
警備員 は 「 ちょっと 待っ て 」 と 私たち を 引き止め た .
When complementizer clauses (CP-THT) are used as a predicate followed by a copula (AX), the extention tag -PRD is added to the label CP-THT.
彼女 と は ただ お茶 を 飲ん で 話 を し た という だけ だ 。
書類 について も 、 オールデイカー の 個人的 な 仕事 も 、 まったく 知ら ない ―― だ そう だ 。
In a raising to object construction, the logical subject of a complement clause (either a complementizer clause or a small clause) is typically marked with particle を and occupies a position in the matrix clause (sibling to the predicate that selects the complement clause). This logical subject is given the extension -DOB1.
In this scenario the complementizer clause is annotated as an IP-SMC dominated by a CP-THT-OB1 (a control environment), and the matrix -DOB1 controls the subject position in the IP-SMC.
自分 は この 言葉 を 面白い と 思っ た 。
警察 は A を 犯人 だ と 断定 し た 。
While the matrix predicate is typically an attitude verb such as 思う or 考える, verbs like する and 扱う can also support this construction.
温度 を 一定 だ と する
In this construction the predicate in the small clause is limited to being either an イ-adjective, a modal with adjective-like inflection, or a copular expression, which must be in the infinitive inflection. In this case, the embedded clause is tagged as IP-SMC-OB1 without further projecting a CP-THT.
お手伝い でき ない の を 心苦しく 思い ます 。
彼ら の 出身 の 村 や 町 に は 、 戦死者 を 名誉 に 思っ て 祈念碑 が 建立 さ れる でしょう 。
These constructions frequently show up with predicates in the spontaneous (ら)れる form or in the potential form.
その 間 が 、 たっぷり 一 時間 は あっ た 様 に 思わ れ ます 。
親不知 の 断崖 を 通過 する 頃 、 車内 の 電燈 と 空 の 明るさ と が 同じ に 感じ られ た 程 、 夕闇 が 迫っ て 来 た 。
These constructions are passivized versions of the raising to object constructions. In such scenarios, the embedded clause is annotated as IP-SMC so that the control relation holds, just as in the active voice scenarios. The raised subject is given the extension -DSBJ.
In this scenario the complementizer clause is annotated as an IP-SMC dominated by a CP-THT-SBJ (a control environment), and the matrix -DSBJ controls the subject position in the IP-SMC.
温度 は 一定 だ と さ れる
招き猫 の 置物 は 、 幸運 を 招き込む と 言わ れ ます .
In this construction the predicate in the small clause is limited to being either an イ-adjective, an adjective-inflecting modal, or a copular expression, which must be in the infinitive inflection. In this case, the embedded clause is tagged as IP-SMC-SBJ, without further projecting a CP-THT.
救心 という 薬 は 味 も 効能 も 仁丹 ぐらい に しか 思わ れ て ない が 、 べラボー に 高価 な ところ が 信仰 さ れる の かも知れない 。
This section explains how expressions specific to the colloquial style are annotated.
The part-of-speech tag INTJ is for interjectives.
Standalone INTJ instances in clauses do not project INTJP layers.
「 ああ 、 王 は 悧巧 だ 。
Only in the following cases do INTJP layers appear:
すると 帽子屋さん は 、 おやおや と 思い まし た 。
In the following attested example, the CP-THT 見えるかいって as a whole serves as an interjectivals.
じゃ 、 私 の 顔 が 見える かい と 一心 に 聞く と 、 見える かい って 、 そら 、 そこ に 、 写っ てる じゃ あり ませ ん か と 、 にこりと 笑っ て 見せ た 。
Neither INTJ nor INTJP instances are conjoined by CONJP layers (see section 3.5). They are just juxtaposed at the same level.
あー 、 はい はい はい 、 特に ね
ああ そうだ 、 こう やれ ば よさ そう だ な 」
FS indicates a false start. This is used particularly in the spoken data when disfluencies such as repetition and restating are observed.
「 ひ 、 ひ 、 人殺しイ …… 」
この 場所 ここ の アリーナ も また ギリシャ の 円形劇場 の よう に 忘我 の 状態 の ため の 場所 です
Ellipses can be found mostly, but not exclusively, in colloquial texts. This secion summarizes the way to annotate them.
For zero pronouns, or omission of instances of NP with core grammatical roles, see section 1.6.4.
When particles that indicate grammatical roles (whether core ones or peripheral ones) of NP are omitted, these NP instances, which are bare, are directly tagged with these roles (e.g., NP-SBJ and NP-OB1).
ご飯 食べ た ?
あれ 、 雨 降っ てる 。
Note that core grammatical roles are preferred to peripheral ones. あいつ in the following sentence can be complemented by both a -SBJ core grammatical role and a -TPC peripheral role. However, becuase of the preference, the -SBJ tag overwrites -TPC.
あいつ 、 どこ 行っ た の かな 。
The next example contains a genuine NP-TPC:
◎ 仮校舎 、 不自由 続く
For discussion of the omission of heads from NP layers, see section 6.3.
For discussion of the omission of heads from IP-ADV layers, see section 6.4.
Sometimes a phrase or a series of phrases is extraposed to the right after predicational syntagms, and sometimes even after sentence final particles. It is counted as an afterthought that complements the gapped sentence body.
When a phrase which is related to the sentence body in any way gets right-extraposed as an afterthought, it is related to the sentence body using *ICH* (see section 1.4). That is, the extraposed phrase leaves an *ICH* trace with an index which is dominated by a node whose label fully specifies the part-of-speech (indicated by the main label) and the overtly annotated grammatical role (indicated by the label extension(s)), if any; the extraposed phrase itself is labelled by the same main label (but without label extensions) and the same index. In example (6.106), the trace is specified as (PP-SBJ *ICH*-1) and the extraposed phrase is tagged as (PP-1 神戸は).
美しい 街 です よ 、 神戸 は 。
来 た 、 来 た 、 バス が 。
/ 生きる 、 全て を 糧 に
When a contraction of two morphemes occurs, then the assignment of a tag is usually made according to the part of speech of the morpheme that determines the function of the whole, typically the second member. Thus, the contracted forms てる (< ている), ちゃう (< てしまう), てく (< ていく), とく (< ておく) are all tagged as VB2. Here are some examples.
僕 、 もう やめ ちゃう けど 、 君 は どう する の 。
ここ に 置い とく よ 。
In general, it is our policy that distinct functional words with homophonous forms be distinguishable from each other by part of speech assignment, structural position, and grammatical function, to the extent that this is reasonably possible. This mainly applies to such things as particles (P), copulas (AX), verbal auxiliaries (AX), and modal elements (MD).
Putting a large variety of functions under one part of speech simplifies analysis but hides important distinctions in function and meaning. As a first step in sorting out the many different uses of basic functional elements, we try to distinguish between a particle and a copular function for items in the following phonological shapes: で, に, と, の.
To this end we identify, for example, a grammatical role particle で in contrast to a verbal suffix particle で and a non-finite copula stem で. The first selects an NP and projects a PP, the second follows a combining stem of a consonant-base verb to specify an inflection, and the last selects a nominal predicate and projects an IP.
There is a limit to the fine-grainedness of the distinctions we make: The role particle (P-ROLE) で will mark indistinguishably:
共通 の ルール を つくる 過程 で は 、 交渉 と 説得 が 行わ れる 。
翌年 、 慶長 20 年 ( 1615 年 ) の 大坂の役 ( 夏の陣 ) 道明寺の戦い で は 後藤基次ら と 戦っ た 。
天正 12 年 ( 1584 年 ) 10 月 に 18 歳 で 家督 を 相続 し 、 伊達家 17 代 を 継承 する 。
うん 、 後 で 返事 し て も 、 いい ?
その 中 の どの わざ の ため に 、 わたし を 石 で 打ち殺そ う と する の か 」 。
太郎 は 風邪 で 学校 を 休ん だ 。
取材 で 行く 西日本 は 東北 と は 食べ物 も 建造物 も 違う 。
この 煙突 は レンガ で 出来 て いる 。
We do not distinguish between these meanings of the grammatical role particle で. For extensions to be attached to PP to differentiate the meanings, see section 1.7.1.
In contrast to the grammatical role particles above, the verbal suffix particle で (an allomorph of て) attaches to onbin forms of voiced velar or nasal consonant base verbs (e.g., 泳ぐ, 読む, etc.): 泳いで, 読んで. This type of で is labelled P-CONN.
また 、 ヘプバーン は 縁起 を 担い で 自身 の ラッキーナンバー で ある 「 55 」 番 の 楽屋 を 求め た 。
Finally, the non-finite copula で, which is labelled AX, appears after an NP-PRD to project an IP-ADV,
ハリウッド黄金時代 に 活躍 し た 女優 で 、 映画界 ならびに ファッション界 の アイコン として 知ら れる 。
or as the stem of analytic copular forms like (AX で) (VB2 ある) and (AX で) (VB2 ござい) (AX ます).
一人 は 、 W君 と いっ て 、 初対面 の 人 で ある 。
There are scenarios where the distinction between a copula and a particle is not easy to make: Descriptions of mode of event and depictive expressions are included in these. At present the practice is to annotate this kind of で following nouns as P-ROLE:
刺身 は 生 で 食べる 。
However, で following an ADJN is annotated as an AX:
崖 を 必死 で よじ登っ た .
Furthermore, in instances of IP-ADV we annotate as AX で following a formal noun (FN), auxiliary (AX) or modal (MD) with a modal or aspectual meaning such as そう, べき, もの, よう, わけ, ごとく, がち, みたい, ところ, っぱなし, etc., regarding the fact that this kind of で is an inflected form of AX such as だ and です which follow these elements in matrix contexts.
お湯 が 煮え立っ た ところ で 、 材料 を いれる .
至極 簡単 明瞭 な 名 だ が 、 この 名前 、 世間 に ザラ に あり そう で 、 滅多 に ない から 不思議 で ある 。
私 は 先生 の 部屋 に 行っ た が 、 入れ違っ た よう で 会え なかっ た .
In many respects the considerations with regard to the phonological form で also hold for に, と, and の. Each of these has a copular function mirrored in its structural position, although the details for each differ.
Particle に marks both core and peripheral grammatical roles. Core particle に marks -SBJ with potential predicates such as できる, 難しい, stative predicates such as 必要だ and 無理だ, cognitive predicates such as 分かる, 見える, 欲しい, etc. It marks -OB1 in a variety of basic predicates, including verbs like 似る, 勝つ, 当たる, 飽きる, etc., イ-adjectives such as 詳しい, 弱い, 等しい, 厳しい, etc., ナ-adjectives such as 必死だ, 反対だ, etc. Furthermore, it marks -OB2 for ditransitive verbs like あげる, 送る, etc. Finally, it marks -CZZ Causee in causative constructions, -DOB1 for predicates selecting instances of IP-SMC-OB1 such as もらう, ほしい, etc., and -LGS in direct passive constructions (See section 2.3.3 for details).
Peripheral role particle に marks temporal and spatial location for non-activity verbs. It also marks the Goal for verbs of motion such as 行く, 着く, etc, the Source for the verbs 借りる, 貰う and いただく, a demoted Agent in unaccusative verbs like 捕まる, 見つかる, 教わる, 弱る, etc., and something similar to an agent with verbs like 満ちる, 溢れる, まみれる. The list is quite long and varied, but we do not distinguish between these meanings of the grammatical role particle に.
Role particle に also marks purpose clauses with verbs of motion, with the purpose clause annotated as a nominalized clause (IP-NMZ; see section 6.16):
わたし は 彼 を 起し に 行く 」 。
に is also analyzed as a copula in some contexts. Primarily heading ADVP instances, we see ADJN followed by (AX に).
また 、 供養墓 が 他 の 大名 など と 同様 に 高野山奥の院 に ある 。
Another context for (AX に) is in small clause constructions:
私 は この 国 に 生まれ落ち た こと を 幸せ に 思う .
怖 そう に も 見え た 。
One more context for copula に is in the small clause complements of verbs する and なる:
嗅覚 とか 味覚 を 刺激 し 、 心 を ほぐし て 豊か に し て くれる 。
ある 日 、 オットー は 借金 の カタ として 自動車 を 回収 する 業者 = レポマン の バッド という 男 と 知り合い 、 自分 も レポマン に なる 事 を 決める 。
Here too are scenarios where the distinction between a copula and a particle is not easy to make: In expressions such as ご褒美に, 担保に, しるしに, etc., に is arguably a copula, but the present practice is to annotate it as a role particle:
メロス の 十六 の 妹 も 、 きょう は 兄 の 代り に 羊群 の 番 を し て い た 。
Resultative clauses also can be analyzed as small clauses, and in this context too に is arguably a copula, but the present practice is to annotate it as a role particle, unless immediately following an ADJN:
どう 考え 、 どう 行動 する か を 判断 する 力 、 すなわち 「 生き 抜く 力 」 を 身 に 付ける ならば 、 原発事故 の マイナス を プラス に 変え られる 。
For と, we distinguish a conjunctive particle (P-CONN), a case particle (P-ROLE), a complementizer particle (P-COMP), and a copula (AX).
As a conjunctive particle, と can conjoin NP instances. Note that instances of と attached to non-final conjuncts head CONJP layers, but the one that is attached to the last conject is ‘buried’ under the whole dominating NP (see section 3.5 for more discussion):
滝役 の 千葉治郎 は 阿部 と 内田 と で 決め た 。
At the IP level, role particle と is primarily comitative. Sometimes the particle marking is determined by the requirements of the verb:
翌年 、 慶長 20 年 ( 1615 年 ) の 大坂の役 ( 夏の陣 ) 道明寺の戦い で は 後藤基次ら と 戦っ た 。
この 年 に は 『 昼下りの情事 』 に も 出演 し て おり 、 ゲーリー・クーパー や モーリス・シュヴァリエ と 共演 し た 。
In other instances the role that the particle marks is peripheral to the predication:
お昼 に 、 卒業生 と 、 なんか 、 ご飯 いっしょ に 食べる ん だ けど
と marks complement clauses of verbs of thought, feeling, and communication:
そっか 、 ごはん 、 いっしょ に 食べ ない かなぁ と 思っ て 。
P-COMP と also marks as a complement (-CMPL) arguments of naming verbs (See section 2.7.3).
日本 で は ヘップバーン と 表記 さ れる こと も 多い 。
と also has a distribution much like infinitive copula に in many respects. It appears at the end of adverbial expressions, and is annotated as (AX と) following ADJN instances that head ADVP layers:
しかし 、 ユダヤ人ら を 恐れ て 、 イエス の こと を 公然 と 口 に する 者 は い なかっ た 。
After common nouns it is still annotated as P-ROLE:
企画段階 で 紆余曲折 を 経 た 本作 は 、 放映開始 さ れ て も 順風満帆 と は 行か なかっ た 。
Likewise, after adverbs which can stand on their own without と, と is annotated as P-ROLE. In contrast, when と appears at the end of an adverbial expression from which no ADV (or ADJN) can be isolated, it remains chunked together with other material in the adverbial (See section 3.4 for details.).
家事 一切 に 関わら ず 、 のんびり と 心 の ぜいたく を する 。
仙台 の 街 を 歩く と 、 妙 に ホッと し て しまう 。
と also appears after instances of ADJN and NP-PRD and in small clause complements to verbs する and なる:
また 、 この 頃 ヘプバーン は 、 アンネ・フランク の 『 アンネの日記 』 を 題材 と し た 舞台作品 と 映画作品 の 両方 へ の 出演依頼 を 受け た 。
礼 に 過ぐれ ば 諂(へつらい) と なる 。
Note that if a copula can be inserted between と and an ADJN or NP-PRD without change of meaning, the particle is analyzed as a complementizer (P-COMP) which projects a CP-THT.
いま 心配 な の は 、 さまざま な 無償 の 支援 を 当然 と 思う 雰囲気 が 生まれる こと だ 。
と attaches to IP-ADV as a conjunctive particle (P-CONN):
太平 の 世 に なる と 美食 を 極める こと に 目的 を 変え て 料理研究 を 続け た 。
の is ambiguous between usages as a case particle (P-ROLE), a so-called formal noun (N/FN), and a copula (AX).
の appears as a grammatical role particle marking an NP as a complement of a N (typically via a possessor/possessee relation or a whole/part relation or a set/member relation):
私 の ゼミ の 卒業生 が 集まっ て 食事会 を する ん です けどもー
の appears as a core grammatical particle (P-ROLE) marking subject NP instances in adnominal clauses:
花子 の 読ん だ 本
隙 の ない 印象 の 政宗 で ある が 、 酒 に は 滅法 弱く 、 酔っ て 失敗 し た 逸話 が いく つ か 残さ れ て いる 。
の can nominalize a clause. This usage is called a formal noun in the traditional study of Japanese Grammar, but is assigned N in our annotation. This is the typical pattern for a pseudocleft clause:
舞台 で イライザ を 演じ て い た の は ジュリー・アンドリュース だっ た が 、 アンドリュース に は 映画出演 の 話 は 来 なかっ た 。
The formal noun の can be also used to nominalize facts and activities:
「 あなた が 裸 で ある の を 、 だれ が 知らせ た の か 。
の appears as a clause-level particle immediately dominated by IP after a finite predicate and before a copula in a main clause (that is, as an element in a のだ construction). In such scenarios, の is labelled FN instead of N:
分から ない の です がー
Sometimes の follows a noun phrase as the adnominal form of a copula (AX の):
もし よかっ たら 、 また 別 の 機会 が あれ ば ・・・ 。
この 行為 は 秀吉 の 派手好み の 性格 を 知っ て の 行い と 伝え られる 。
岩手 が 多い 傾向 は 冬期間 を 除き 、 震災発生 から 1 年 半 後 の 12 年 9 月 まで 続い た 。
See section 4.23 for discussion on the criteria for distinguishing の as particle from copula.
There is a polysemous verb form that is formed by adding れる/られる to the relevant combining stem of a verb: られる is added to the i/e-combining stem for vowel-ending verbs, and こ for 来る, and れる is added to the a-stem of consonant-ending verbs. These verb forms are ambiguous between the direct passive, the indirect passive, and the honorific usages. Moreover, for vowel base verbs and the irregular verb 来る, the spontaneous usage and the potential usages are also formed in the same way, making れる/られる end up eventually being five-way ambiguous for these verbs. This section discusses the extent to which we distinguish these usages, and the means we adopt to do it.
In a direct passive construction, semantically, a -SBJ has a Patient role with respect to the core predicate (VB) (corresponding to either an -OB1 or an -OB2 in a construction in active voice) and -LGS has an Agent role with respect to the core predicate (VB) (corresponding to a -SBJ in a construction in active voice). Thus, in a two-place predication, if A-が Bに VB1(ら)れる is true, then B-が Aを VB1 is also true.
The verbal auxiliary れる/られる in the direct passive construction is assigned the label PASS, and ‘NP に / から / によって’ is an -LGS (a logical subject).
太郎 は 朝早く 、 友人 に 電話 で 起こさ れ た 。
See section 6.9 for more discussion.
The indirect passive is a valence-increasing construction which adds an affected subject -SBJ at the same time that it demotes the logical subject of the core predicate to -LGS. The internal arguments of the core predicate preserve their positions and grammatical roles, and the passive morpheme れる/られる is tagged as PASS2.
太郎 が 雨 に 降ら れ た 。
鈴木さん は 昨夜 、 一 晩中 子供 に 泣か れ て 困っ た 。
Note that while some grammars identify scenarios where a Possessor is -SBJ and a Possessee is -OB1 in a passive construction as being a special case of direct passive with an underlying Source argument (possessor passive), most syntactic arguments for that analysis have counterevidence. Such scenarios are treated as indirect passives here.
太郎 が 泥棒 に 財布 を 盗ま れ た 。
そう で ない もの は 塵塚 に 捨て られ 、 存在 を さえ 否定 さ れ た 。
Note: Whether or not the subject -SBJ is adversely affected is not at issue in identifying this construction. The crucial question is whether the construction is valence-increasing. The presence of an を-marked -OB1 is not sufficient to decide the matter, either. If the predicate is three-place and the subject is an underlying -OB2 in a corresponding active voice construction, this is a case of direct passive rather than indirect.
In the spontaneous usage of れる/られる, main verbs are limited to those that mean cognition, thinking, or feeling. This type of れる/られる is labelled VB2.
An experiencer -SBJ (marked with either に, は, or には) is the logical subject of the core predicate, while the internal argument of the core predicate is marked with が but retains its -OB1/-DOB1 label. The auxiliary morpheme is labelled VB2.
そういった 未来 が 予想 さ れ ます
私 に は その 話 が 不思議 に 思わ れ た 。
そういう 事 も 今 に なっ て 考え て 見る と 、 甚だ 奇怪 に 感じ られる の で ある 。
There exist verbs which mean spontaneity or the likes in nature, some being derived and lexicalized from consonant base verbs (their potential form or passive form), and some being the unaccusative members of transitivity pairs:
They are treated as independent verbs and are labelled VB.
The potential form of vowel-ending verbs (and the irregular verb 来る) is formed in the same way as the direct passive, indirect passive, spontaneous, and honorific forms, adding られる to the relevant combining stem (viz. the i/e-combining stem for vowel base verbs, and こ for 来る): 見られる, 食べられる, こられる. An alternative colloquial form adds れる to these: 見れる, 食べれる, 来れる. In both cases the added morpheme is labelled VB2.
This contrasts with the potential form for consonant-ending verbs, formed on an e-combining stem: 走れる. This is left unanalyzed as a whole VB.
When できる is used as a light verb following a verbal noun it is labelled VB0. Generally, the potential form for する is provided by suppletion: できる.
えっとー 、 そういった 感じ で も 参加 できる ん です かー ?
For two-place predicates which can take -OB1 with を, core particle marking in potential constructions can follow any of three patterns.
むしろ 、 自分 が これ まで 実際 に この かぼそい 脚 で 身体 を ひきずっ て こ られ た こと が 不自然 に 思わ れ た 。
どうして こんな 妹 が かせぐ こと が できる だろう か 。
では 私たち に 何 が できる の でしょう か ?
The status of the subjects in potential constructions is a bit complex. Their denotations typically have animate, sentient, human features. A (NP-SBJ *arb*) subject in a sentence like 一回目で自転車にうまく乗れないのは誰だって一緒だ is perfectly normal. But a sentence like この山が登れない is odd unless a retrievable subject (‘we, they, everyone,’ etc.) is available for the interpretation.
この 山 が 登れ ない 。
In the honorific usage of the れる/られる form, the core grammatical roles and labels are retained unchanged. れる/られる is tagged as VB2.
鈴木先生 は その 問題 を 詳しく 調べ られ た 。
For other honorific verb forms, see section 4.6.
There are mainly three labels for the element よう.
This section discusses (MD よう) and (N よう).
The evidential ending ようだ ‘It seems that...’ which follows the past and the non-past form of verbs and イ-adjectives, [ナ-adjective + な/だった], and [noun + だった] is tagged as MD.
この 事実 から 判断 する と 、 あなた の 見方 は 正しく ない よう だ 。
In the NPCMJ, this evidential use of よう is distinguished from the simile use of よう, which is discussed below. Inferential adverbial phrases such as どうやら, どうも can be used to detect the former: if such adverbial phrases can be inserted without much changing the meaning of the entire clause, the よう in question is evidential.
どう も 太郎 は この こと を 知っ て いる よう だ 。
In ending a non-final conjunct in a clasual coordination at main clause level, evidential よう is followed by the non-finite copula で.
私 は 先生 の 部屋 に 行っ た が 、 入れ違っ た よう で 会え なかっ た .
When よう is used to express a simile meaning, よう is tagged as an N which takes an IP-EMB. Adverbial phrases like まるで, あたかも are typical indicators of よう's simile use (note that these adverbial phrases are located in the same clause as the copula, which is supported by the fact that they can occur with copulae even in the absence of よう).
頬冠り に 尻端折り 、 草履 は 懐中 へ 忍ばせ た もの か 、 そこ だけ ピクリ と 脹れ て いる の が 蛇 が 蛙 を 呑ん だ よう だ 。
まるで 酔っ て いる よう だ 。
Also, when よう takes an IP-EMB complement and itself modifies other words (in configurations such as ような), it is tagged as N.
あいつ は 俺 に 感謝 する どころか 、 おれ の 顔 に 泥 を ぬる よう な こと ばかり し て くれる 。
In the complementizer complex [よう + に] used with some cognitive verbs, too, よう is tagged as N, with に projecting a small clause (IP-SMC-SBJ or IP-SMC-OB1):
あたかも 日本 の 製造業 全体 の 競争力 は 強い か の よう に 思わ れ て き た 。
部屋 の 闇 の 中 で 僕 は 今 また 麻痺 し た 人 の 重たい 灰色 の 顔 を 見 た よう に 思っ た 。
In addition, there are other verbs whose complement is marked by ように such as 振る舞う. One common feature of cognitive verbs and these verbs is that the simile-indicating adverbs まるで, あたかも can modify ように. Therefore, the same annotation strategy is appropriate here, too. The use of ようにする, ようになる which allows the modification by these adverbs should also be annotated in the same way.
彼 は あたかも 責任者 の よう に 振る舞っ て いる 。
背中 が 棒 の よう に なっ た 。
Furthermore, sometimes clauses introduced by ように that allow the insertion of the adverbs まるで, あたかも appear not as an argument of verbs but as a modifier of verbs' manners or situations expressed by the entire sentence. In these cases, よう is tagged as N and に as AX which projects IP-ADV-SCON. Note that manner modification intended here is different from that discussed above. The crucial criterium for this use is whether the modification by まるで, あたかも is possible.
リニアモーターカー は あたかも 氷 の 上 を すべる か の よう に 走る 。
ザムザ夫人 と グレーテ と は また 書き つづけ よう と する か の よう に 手紙 に かがみこん だ 。
As a consequence of the IP-SMC analysis, we can have two clauses: an IP-EMB embedded by よう and an IP-SMC headed by に. This makes it possible to capture cases where the source of impression (the controller of the IP-SMC) isn't identical to the subject of the IP-EMB. For example, the following sentence, intuitively, describes the impression about 丸善 (the controller of the IP-SMC), whose content is that not 丸善, but the speaker (the subject of the IP-EMB), can easily enter the store.
平常 あんな に 避け て い た 丸善 が その 時 の 私 に は やすやす と 入れる よう に 思え た 。
However, by adopting this approach, we always have to care about what the controller of the IP-SMC is, which is not always clear, as in the following sentence (Clearly, 私には in the matrix clause is not the controller). For this case, *pro* is introduced as the controller, but we might as well use *exp* for it has no plausible antecedent, though there is no existing example in which *exp* is dominated by an NP-OB1.
しかし 、 私 に は 、 庶民 の ごく 素朴 な 想像力 において は 、 神道 の 原始的 な 概念 と 、 仏教 の 教え で ある 魂 の 裁き という はるか に 明白 な 教義 と の 間 に は 、 真 に 合致 する もの は 明らか に ない よう に 思える の で ある 。
This strategy might gain support from cases where the IP-EMB complement to よう has some quantificational element scoping over its subject (i.e., the subject's referents vary depending on individuals/times/worlds/etc. quantified over by that element). It is not clear how to express this scope relation if we locate the subject in the same clause as the embedding predicate. The following is a constructed example:
Such a scoping possibility might constitute a real problem in other cases involving an IP-SMC, whose subject must be always controlled by another element above it, but without any lower positions in which to locate the subject in order to represent the appropriate scope relation. Sometimes, morphological marking by を even requires the subject of the IP-SMC to be in the same clause as the embedding predicate. The following is an example from Google.
よう(だ) following [NP + の] is always labelled N irrespective of whether it is used as the evidential よう or as the simile よう.
太郎 は まだ 子供 の よう だ 。
この 絵 は 写実的 で 、 写真 の よう だ 。
The same is true of よう(だ) following a determiner such as この, その, あの, and どの.
この よう な こと を 知っ て いる 人 は 、 もはや い ない 。